The Placebo Effect

[HansMustermann]

Since all pain is self-reported to begin with, how can you actually know they were in pain in the first place? When it comes to the study of pain, all we have to work with is self-reporting. Might as well ask if the "green" that you see happens to be the same color as the "green" that I see.

Quoth Dara O'Briain, "Science knows it doesn't know everything; otherwise, it'd stop. But just because science doesn't know everything doesn't mean you can fill in the gaps with whatever fairy tale most appeals to you."

Now I'm not saying that the study of pain is at the fairy tale level, but still, it doesn't mean that LACKING better data necessarily makes a line of thinking correct.

 

[theprestige]

I beg you, don't go there.

You are now talking about qualia, we are deep in Philosophy territory.

We were already deep in philosophy territory when Hans asked how we can use self-reporting as a metric for change in something for which we only have self-reporting to work with. If Hans doesn't want this to degenerate in navel-gazing about the qualia of pain, he needs to get back to practical applications.

 

[HansMustermann]

In all the examples you are referring to, there is a distinction between the experience and the underlying thing that is being experienced. And you keep pointing out that the underlying thing being experienced isn't any different, even if the experience is different because people have fooled themselves.

But pain IS the experience. The underlying thing being experienced isn't pain, but physical damage. And yes, we know placebos don't affect physical damage to the body. That doesn't mean they don't affect the experience. If you think it affects the experience, guess what? It does. By definition.



You are being asked to believe that people's perceptions of pain IS the pain, that pain isn't anything else. The proper comparison to all your examples would not be asking people how much pain they suffered, but how badly their body was damaged. And yeah, people won't get that right. But since the challenge we're talking about isn't fixing damage but alleviating pain, then their experience is rightly front and center.

1. No. It goes beyond that. I'm asked to believe that people can actually make an accurate comparison for such a sensory signal / qualia, even though they provably CAN'T do it for any other senses / qualia.

Regardless of what other processing is done along the way, what feedback loops might exist, etc. SOME value / qualia reaches the final destination on the brain. It may be processed to hell and back, it may not reflect the input particularly well any more, but it's still a value at the end of that processing chain. Whether it's pain, sound, the sense of time, or whatever, but it's how they perceive that pain or sound or time or whatever RIGHT NOW.

And we KNOW that people are utter crap at comparing such non-quantized values even to the value from 5 minutes ago. The guy on the HW Central forums who was comparing MP3 sound from different hard drives, was essentially crap at comparing how it sounded 1 minute ago, when he tried listening to the same MP3, on the same computer, on the same sound card, on the same MP3 player software, on the same headphones, just copied to a different hard drive.

Hell, we know that stuff like cognitive dissonance and generally how engrams work can actually CHANGE that memory.

So the same people who are utter crap at comparing the non-quantized qualia for sound, time, lag, and generally EVERYTHING, even to something from 5 minutes ago... I'm supposed to believe that they're 100% accurate at comparing their current pain to the pain from last week, before they were given the placebo pills


2. That is all good and fine, but please explain in simple words why couldn't the exact same reasoning be applied to all the other examples I gave, if we didn't have the means to actually measure that the perception is wrong.

I mean, take sound for example. Let's say we didn't really know much about how the brain figures out the frequencies and all (you only need to go back a couple of centuries for that.) Equally you could say that the music is the experience, there is no objective way to perceive it, the nerves signal both ways, the brain doesn't get the raw frequencies, etc. In fact, I'm sure some "audiophile" out there did the exact same argument when shown on an oscilloscope that his 1000$ wooden volume knob doesn't actually make his stereo reproduce high frequencies better. (And yes, there are people buying just that kind of nonsense and arguing that it makes their music sound better.)

But in fact, we know that the frequencies are already detected and separated at the cochlea, and there are no nerves to tell it to send more of the high frequencies.

 

[theprestige]

Quoth Dara O'Briain, "Science knows it doesn't know everything; otherwise, it'd stop. But just because science doesn't know everything doesn't mean you can fill in the gaps with whatever fairy tale most appeals to you."

Now I'm not saying that the study of pain is at the fairy tale level, but still, it doesn't mean that LACKING better data necessarily makes a line of thinking correct.

Nobody's talking about it being necessarily correct. We're saying it's the only line of thinking currently available. I'm sure the moment someone comes up with a better metric for pain mitigation that self-reporting, the entire medical community will stand up, clap enthusiastically, then then start using it almost exclusively. In the mean time, "but how do we know if self-reporting is really accurate?" is kind of a pointless waste of time. We know self-reporting has limitations. We know it's not as reliable as some other metrics. It is what it is.

 

[HansMustermann]

We were already deep in philosophy territory when Hans asked how we can use self-reporting as a metric for change in something for which we only have self-reporting to work with. If Hans doesn't want this to degenerate in navel-gazing about the qualia of pain, he needs to get back to practical applications.

So you went into a discussion about subjective pain and generally subjective reporting of how one subjectively feels better after a placebo, expecting it to NOT involve talking about how it's subjective? REALLY?

Well, anyway, if you're not interested in this particular topic, you have a back button up there, you can feel free to use it. I have no duty to keep everything dumbed down to whatever level passes your approval

 

[theprestige]

1. No. It goes beyond that. I'm asked to believe that people can actually make an accurate comparison for such a sensory signal / qualia, even though they provably CAN'T do it for any other senses / qualia.

No. You're being asked to ACCEPT - or at least ACKNOWLEDGE - that given the nature of pain, self-reporting is the best metric we have, for all its known limitations. [/QUOTE]

 

[HansMustermann]

No. You're being asked to ACCEPT - or at least ACKNOWLEDGE - that given the nature of pain, self-reporting is the best metric we have, for all its known limitations.

Except that is not what was being said, or what such studies generally claim.

And yours is a nonsense argument anyway. If some way of measuring things is inaccurate, then yes, it's valid to discuss exactly what is the margin of error, and thus how accurate or inaccurate one can expect the results to be. It's in fact how science works. Not having anything better is not precluding any of that.

So to use your own words, you're being asked to ACCEPT -- or at least ACKNOWLEDGE -- that not all conversations are with you, nor do they have to conform to your level of misunderstanding.

 

[theprestige]

So you went into a discussion about subjective pain and generally subjective reporting of how one subjectively feels better after a placebo, expecting it to NOT involve talking about how it's subjective? REALLY?

I'm saying the subjectivity is a red herring. Your objection is pointless because there aren't any other options for measuring pain and pain mitigation.

Well, anyway, if you're not interested in this particular topic, you have a back button up there, you can feel free to use it. I have no duty to keep everything dumbed down to whatever level passes your approval : p

I'm interested in the topic. I'm fascinated by your idea that we shouldn't measure pain via self-reporting, even though we don't have any better tools, because it's not an ideal tool.

Tell me more. Tell me what you recommend, as far as the study of pain and its treatment goes. Should we just ignore the entire field, because we can't measure it as reliably as you would like? Should we stop prescribing pain medication at all, because who knows if the patient really is feeling pain in the first place? Should we dismiss the patient who asks for a different prescription, since we have no idea if they're right about their current prescription not working?

What is the practical application, for medicine and pain research, of your objection about the reliability of self-reporting pain?

For me, your objection seems like a dead end. The only way we know someone is in pain to begin with is because they claim to be in pain. If you're going to accept the a priori premise that their self-reported claim is true and indicates a legitimate need for treatment, then it seems silly and pointless to question whether their self-reported claim about the efficacy of the treatment is also true.

What exactly is the practical application you are proposing, from the premise that pain is qualia? What practical application could there reasonably be, other than, "qualia is what we have, so qualia is what we're going to work with. Be mindful of its implications, and do the best you can"?

 

[Ziggurat]

1. No. It goes beyond that. I'm asked to believe that people can actually make an accurate comparison for such a sensory signal / qualia, even though they provably CAN'T do it for any other senses / qualia.

Again, no. Every disproof you've appealed to is comparing people's senses to something external, and showing that those senses don't accurately represent the objective reality of that external thing.

And yeah, our senses are so mediated that they are not accurate or reliable representations of those external realities at a fine level. But NONE of that is comparable here, because (again) we aren't talking about anything external. EVERYTHING is internal. And pain specifically is entirely within the brain. There is no such thing as pain outside of the brain. Again, the equivalent of all your examples is not whether something reduces pain, but whether something reduces damage. And yeah, people will get that wrong, pretty reliably. But that's not the issue here.

2. That is all good and fine, but please explain in simple words why couldn't the exact same reasoning be applied to all the other examples I gave, if we didn't have the means to actually measure that the perception is wrong.

Because in the case of pain, the perception is the thing. In every example you have given, the perception is NOT the thing. In all your examples, we care about something external, and people are using their perception to evaluate it. In the case of pain, ONLY the perception matters, because again, that's what pain is. Pain is perception of damage, but we aren't talking about what treats the damage, only about what alters the perception.

 

[HansMustermann]

I'm saying the subjectivity is a red herring.

It's not if we're discussing the limits and error margin of subjective measurements.

Your objection is pointless because there aren't any other options for measuring pain and pain mitigation.

And again: we still discuss the error bar in any scientific endeavour, even if we don't have better tools. Unless you've slept in school, even if the best tool you have to measure, say, a school level interference experiment is a ruler marked in sixteenths of an inch, you still put some thought into what level of accuracy you have for the measured wavelength. The fact that you don't have a better one doesn't absolve you from that.

And yes, that includes some thought into whether the tools you have are actually adequate for the job. Like if the only tool you have around for measuring heights is a barometer, it may be enough to measure a mountain height, but not enough to measure your house's height.

Or even in subjective stuff, like clinical studies (which is what we're talking about) or sociology or really anything else, it's actually important to know if your results are (A) compensating for other variables and known effects, and (B) statistically significant. The latter basically meaning the probability to be wrong. If all you can do is an imperfect job in those aspects, it is important to be aware of that, rather than pretend that you must ignore the limitations because it's the best you can do.

But basically all you show here is that you don't even understand what's being discussed or why. And you just do your usual thing of going around and complaining that people put way too much thought into this or that, or actually make judgments about this or that, instead of keeping it at whatever level of simplicity and certainty you'd like.

I'm interested in the topic. I'm fascinated by your idea that we shouldn't measure pain via self-reporting, even though we don't have any better tools, because it's not an ideal tool.

And I'm fascinated what kind of confusion of mind makes you think anyone gives a flip about what you want the talks in various threads to be more like. Maybe going around demanding attention in every talk worked with your mommy, but frankly, you should have grown up and out of that by now.

 

[theprestige]

It's not if we're discussing the limits and error margin of subjective measurements.

Is that what you want to talk about? The error margins on pain reporting? Then go ahead and talk about it. You put a lot of effort into saying this is what you want to talk about, and no effort at all into actually talking about it. Here, for example:

And again: we still discuss the error bar in any scientific endeavour, even if we don't have better tools. Unless you've slept in school, even if the best tool you have to measure, say, a school level interference experiment is a ruler marked in sixteenths of an inch, you still put some thought into what level of accuracy you have for the measured wavelength. The fact that you don't have a better one doesn't absolve you from that.

And yes, that includes some thought into whether the tools you have are actually adequate for the job. Like if the only tool you have around for measuring heights is a barometer, it may be enough to measure a mountain height, but not enough to measure your house's height.

Or even in subjective stuff, like clinical studies (which is what we're talking about) or sociology or really anything else, it's actually important to know if your results are (A) compensating for other variables and known effects, and (B) statistically significant. The latter basically meaning the probability to be wrong. If all you can do is an imperfect job in those aspects, it is important to be aware of that, rather than pretend that you must ignore the limitations because it's the best you can do.

But basically all you show here is that you don't even understand what's being discussed or why. And you just do your usual thing of going around and complaining that people put way too much thought into this or that, or actually make judgments about this or that, instead of keeping it at whatever level of simplicity and certainty you'd like.



And I'm fascinated what kind of confusion of mind makes you think anyone gives a flip about what you want the talks in various threads to be more like. Maybe going around demanding attention in every talk worked with your mommy, but frankly, you should have grown up and out of that by now.

Walk me through it. What's the practical application? Someone comes to you reporting pain qualia. What should you do? Offer them some aspirin? Tell them you'll need to see some error bars? What? If they take the aspirin and report the qualia has subsided, what should you do? Accept their claim? Dismiss their claim? Ask for error bars?

 

[HansMustermann]

Pain is perception of damage, but we aren't talking about what treats the damage, only about what alters the perception.

Err.... No, not really. Pain is a signal that comes down a nerve. It can mean damage, or it can mean no damage at all. Not all pain receptors respond to cells dying. In fact some have nothing at all to do with that.

As a trivial example, your thermal pain receptors (nociceptors) are literally responding not just to the death of any cells around, but also to temperature. And your threshold is probably around 44–45 °C in the upper direction if you're an average human, which is well below the level where cells are dying right now. It's more to prevent actual damage than to actually measure damage. But anyway, you can feel that kind of pain without ANY actual damage happening.

They're even "misused" as taste receptors in your mouth. E.g., capsacin literally binds to the the temperature receptors.

Other nociceptors respond literally to pressure. Even if no cell around has actually died, above a certain pressure level, you'll get a pain signal.

Yet other nociceptors (the "silent nociceptors") are only activated when there is an inflammation (including due to cell death) or being stimulated too often or such. Unless the proteins associated with an inflammation are there to activate the receptors that need an inflammation, they just don't trigger. Joints for example contain a LOT of these. So for example the same pressure in the joint might hurt like heck if it's inflamed, but not hurt at all if it's healthy. Again, the reason seems to be more to prevent further damage, than respond to actual damage happening right now.

Yet others respond to chemicals in the tissue, so they work kinda like the taste receptors in the tongue. A lot of those chemicals is how the ones that respond to tissue damage do so (and in fact, there are several receptors for several of these chemicals released when repairing tissue damage) but some respond to specific substances like lactic acid (think: muscle sores) or just to the PH being above or below a limit. Tissue damage may or may not be involved, but sometimes it's more about keeping you from getting to the level that actually destroys tissue, than recording existing tissue death. The lactic acid receptors are the prime example for that: they actually respond to lesser concentrations than would actually kill the muscle cell, to keep you from getting to that level.

Etc.

So essentially the objection that 'yeah, but pain doesn't just mean injury', yeah, that's nothing new.

It's still a signal that comes up a nerve, and gets processed SOMEHOW before becoming a subjective sensation that you have to compare to another subjective sensation. No different from any other, like sound, temperature or whatever. If there are effects skewing the comparisons done for the others, I want to know if they're taken into account for this one too.


Also, yes, I know about the inhibition mechanisms but even those aren't as clear cut mood-over-matter as you seem to think. A LOT of it happens entirely in the spinal cord and really has nothing to do with your mood or stress level, because those nerves don't know anything about that. Yes, some of the suppressive signal comes from above in the form of serotonin, norepinepherine and dopamine, but a lot of it is entirely local in the spinal cord and has nothing to do with those or your mood or stress level.

And even those three have more to do with what you're doing at the moment, rather than what you think about the pill you took three hours ago.

E.g., dopamine has more to do with motivational salience, i.e., more of a "yeah, do that" or most often a "keep doing that" signal, as in whatever you were doing at the moment, than having anything to do with what you think about the pill you took three hours ago. Or really with ANYTHING you were doing three hours ago. As I was saying, it's about regulating what you're doing right now.

So it explains things like that you might notice less pain when you're focused on, say, a computer game, but buggerall to do with whether you think the pills you took this morning are working or not


But anyway, yeah, sure, some processing is involved between a pressure receptor firing in your hip and your brain getting a pain value. But it still gets A value at the end of the day. Is it any better at comparing it with the value from yesterday than for any other signal? Is it compensating for all the other variables involved? Having more variables along the way just means more room for subjective confirmation, rather than making it objective.

 

[theprestige]

How do you even put error bars on something like self-reporting pain? "There's an X per cent chance that this person is not experiencing the amount of pain they say they're experiencing." How does that work?

 

[HansMustermann]

Is that what you want to talk about? The error margins on pain reporting? Then go ahead and talk about it. You put a lot of effort into saying this is what you want to talk about, and no effort at all into actually talking about it. Here, for example:



Walk me through it. What's the practical application? Someone comes to you reporting pain qualia. What should you do? Offer them some aspirin? Tell them you'll need to see some error bars? What? If they take the aspirin and report the qualia has subsided, what should you do? Accept their claim? Dismiss their claim? Ask for error bars?

What I'm saying is: that's a fundamentally stupid attitude. A lot of the progress we've done on that topic comes from trying to see how we can do or measure things better.

E.g., that you have Ibuprofen (which btw, celebrates its 50 year anniversary this year) as an alternative, yeah, that comes from trying to find a safer alternative to aspirin. Which included coming up with such stuff as using a radioactive form of the drugs they were testing in lab rats, to see where it ends up in the body. Such as whether it accumulates in the liver.

The fact that now you can also use it as an anti-inflammatory came from people actually trying to figure what it does to the tissue, rather than "just take one and tell me if you feel better." The actions of prostaglandins in mediating and regulating inflammation wouldn't even be known until a decade later, and yeah, it came about because people were trying hard to measure more than before. Including seeing exactly with what of your body's chemistry it actually interacts.

And that was only the start. Currently there is actually a lot of computational research in simulating protein interactions, to develop new analgesics, by knowing what you want them to bind with and what you don't want them to bind with. A lot of it boiling down to something that binds with certain opioid receptors (those in the actual pain receptors), but not with the other receptors that traditional opioids bind with.

THAT is what comes out of looking at what actual signals are involved and affected, rather than keeping it at your favourite level of "derp, just take an aspirin."

 

[HansMustermann]

How do you even put error bars on something like self-reporting pain? "There's an X per cent chance that this person is not experiencing the amount of pain they say they're experiencing." How does that work?

Well, you can start by looking at what amount of error do you see in other subjective things, that you can measure. E.g., how many people say they hear more bass with headphones A than headphones B, when they're actually the same labelled differently, but you told them that B is the extra-bass version. Does the percentage of people reporting placebo effects exceed that, and by how much?

 

[theprestige]

Well, you can start by looking at what amount of error do you see in other subjective things, that you can measure. E.g., how many people say they hear more bass with headphones A than headphones B, when they're actually the same labelled differently. Does the percentage of people reporting placebo effects exceed that, and by how much?

But pain is qualia. Unlike your audio channel test, there's no objective standard to measure against.

And you can't just say that X% of people misreport their audio channel, so X% are probably misreporting placebo effects on pain. What would misreporting their pain even mean? They're lying about how much pain they're experiencing? They're mistaken about how much pain they're experiencing? That makes no sense.

"You say you're imagining less pain, but there's an X% chance that the pain you're imagining is actually more pain."

 

[HansMustermann]

People do lie all the time about all kinds of things, you know?

I just told you why they have to randomize polls, for example. If you ask ANY yes/no question without randomizing it, we KNOW there'll be a skew towards "yes". You could ask a random group A to answer "should we continue the war" and a group B "should we stop the war" and you'll find that the numbers aren't mirrored. In fact if you add the percentage of "yes" in group A to the "yes" percentage for group B (which should mirror the "no" option in group A), you get over 100% every time. That's why you don't do either in an actual poll, but rather randomize whether someone gets the question in the positive or negative form.

There's also a KNOWN effect where if there's any hint that one option is more acceptable or normal for any reason whatsoever, you get a skew towards it. That's why serious polling companies actually put a lot of effort into making it as neutral sounding as humanly possible.

And it's not just about experiences they've been told about in advance, so some kind of confirmation bias or placebo effect could actually affect the actual experience when it happens. If you ask people what they thought about Angela Merkel's speech about the floods last month, again, you see the above effects. They didn't know they'll be polled when they actually had the experience. So your question phrasing can't have affected the ACTUAL experience they had at the time, without involving time travel. But if you ask them about it after a week, provably you get more "yes" if you ask "did you approve of AM's speech" than "no" if you ask "did you disapprove of AM's speech".

So dunno, considering that as far as we know it applies to EVERYTHING, do you see any reason why only for pain, nah, people totally wouldn't skew the numbers similarly?

 

[xjx388]

Well, you can start by looking at what amount of error do you see in other subjective things, that you can measure. E.g., how many people say they hear more bass with headphones A than headphones B, when they're actually the same labelled differently, but you told them that B is the extra-bass version. Does the percentage of people reporting placebo effects exceed that, and by how much?

The problem with pain perception and measurement is understood. Maybe one day, we’ll have more objective measurements but I kinda doubt it. Some people, due to genetic quirks or other factors, have high pain tolerance or don’t feel pain when they really should. I don’t see how some kind of normalized pain measurement standard can be developed. “Patient reports 8/10 on the Pain Scale but the Mustermann Test says it is actually a 2 . . . they must be lying.” That’s not something I can envision happening.

I think it’s far more likely we will gain a greater understanding of the psychological factors at play and take a multifaceted approach to pain management.

 

[arthwollipot]

I'm keeping an eye out and will post any additional articles if and when Mike puts them up on the site.

 

[HansMustermann]

The problem with pain perception and measurement is understood. Maybe one day, we’ll have more objective measurements but I kinda doubt it. Some people, due to genetic quirks or other factors, have high pain tolerance or don’t feel pain when they really should. I don’t see how some kind of normalized pain measurement standard can be developed. “Patient reports 8/10 on the Pain Scale but the Mustermann Test says it is actually a 2 . . . they must be lying.” That’s not something I can envision happening.

At the level of one patient? No. But I think one should be aware of the factors that skew the aggregate results when talking about a group statistic where the whole point is whether something actually works.

After all, that's the whole POINT of comparing it to placebo in other clinical trials. You don't know if ONE specific test subject you use to test if the new antibiotic pill worked really has a better immune system than average, or had a remission for other reasons, and the point isn't to diagnose that for that one patient in the first place. But you know that x% in the control group recovered while taking just pure lactic acid pills, and you compare if the y% for your actual antibiotic is higher than that. You know that the results are skewed by x% so you subtract that percent from your actual results.

I think it’s far more likely we will gain a greater understanding of the psychological factors at play and take a multifaceted approach to pain management.

Here's to hope.