The most important study on the placebo effect is Hróbjartsson and Gøtzsche’s Is The Placebo Powerless?, updated three years later by a systematic review and seven years later with a Cochrane review. All three looked at studies comparing a real drug, a placebo drug, and no drug (by the third, over 200 such studies) – and, in general, found little benefit of the placebo drug over no drug at all. There were some possible minor placebo effects in a few isolated conditions – mostly pain – but overall H&G concluded that the placebo effect was clinically insignificant. Despite a few half-hearted tries, no one has been able to produce much evidence they’re wrong. This is kind of surprising, since everyone has been obsessing over placebos and saying they’re super-important for the past fifty years.
What happened? Probably placebo effects rode on the coattails of a more important issue, regression to the mean. That is, most sick people get better eventually. This is true both for diseases like colds that naturally go away, and for diseases like depression that come in episodes which remit for a few months or years until the next relapse. People go to the doctor during times of extreme crisis, when they’re most sick. So no matter what happens, most of them will probably get better pretty quickly.
In the very old days, nobody thought of this, so all their experiments were hopelessly confounded. Then people started adding placebo groups, this successfully controlled for not just placebo effect but regression to the mean, and so people noticed their studies were much better. They called the whole thing “placebo effect” when in fact there was no way to tell without further study how much was real placebo effect and how much was just regression to the mean. If we believe H&G, it’s pretty much all just regression to the mean, and placebo was a big red herring.
The rare exceptions are pain and a few other minor conditions. From H&G #3:
We found an effect on pain, SMD -0.28 (95% CI -0.36 to -0.19)); nausea, SMD -0.25 (-0.46 to -0.04)), asthma (-0.35 (-0.70 to -0.01)), and phobia (SMD -0.63 (95% CI -1.17 to -0.08)). The effect on pain was very variable, also among trials with low risk of bias. Four similarly-designed acupuncture trials conducted by an overlapping group of authors reported large effects (SMD -0.68 (-0.85 to -0.50)) whereas three other pain trials reported low or no effect (SMD -0.13 (-0.28 to 0.03)). The pooled effect on nausea was small, but consistent. The effects on phobia and asthma were very uncertain due to high risk of bias.
So the acupuncture trials seem to do pretty well. This probably isn’t because acupuncture works – some experiments have found sham acupuncture works equally well. It could be because acupuncture researchers have flexible research ethics. But Kamper & Williams speculate that acupuncture does well because it’s an optimized placebo. Normal placebos are just some boring little pill that researchers give because it’s the same shape as whatever they really want to give. Acupuncture – assuming that it doesn’t work – has been tailored over thousands of years to be as effective a pain-relieving placebo as possible. Maybe there’s some deep psychological reason why having needles in your skin intuitively feels like the sort of thing that should alleviate pain.
I want to add my own experience here, which is that occasionally I see extraordinary and obvious cases of the placebo effect. I once had a patient who was shaking from head to toe with anxiety tell me she felt completely better the moment she swallowed a pill, before there was any chance she could have absorbed the minutest fraction of it. You’re going to tell me “Oh, sure, but anxiety’s just in your head anyway” – but anxiety was one of the medical conditions that H&G included in their analysis. Plausibly they studied chronic anxiety, and pills are less good chronically than they are at aborting a specific anxiety attack the first time you take them. Or maybe her anxiety was somehow related to a phobia, one of the conditions H&G find some evidence in support of a placebo for. (Really? Phobia but not anxiety? Whatever.)
Surfing Uncertainty had the the best explanation of the placebo effect I’ve seen. Perceiving the world directly at every moment is too computationally intensive, so instead the brain guesses what the the world is like and uses perception to check and correct its guesses. In a high-bandwidth system like vision, guesses are corrected very quickly and you end up very accurate (except for weird things like ignoring when the word “the” is twice in a row, like it’s been several times in this paragraph already without you noticing). In a low-bandwidth system like pain perception, the original guess plays a pretty big role, with real perception only modulating it to a limited degree (consider phantom limb pain, where the brain guesses that an arm that isn’t there hurts, and nothing can convince it otherwise). Well, if you just saw a truck run over your foot, you have a pretty strong guess that you’re having foot pain. And if you just got a bunch of morphine, you have a pretty strong guess that your pain is better. The real sense-data can modulate it in a Bayesian way, but the sense-data is so noisy that it won’t be weighted highly enough to replace the guess completely.
If this is true, placebo should be strongest in subjective perceptions of conditions sent to the brain through low-bandwidth relays. That covers H&G’s pain and nausea. It doesn’t cover asthma and phobias quite as well, though I wonder if “asthma” is measured as subjective sensation of breathing difficulty.
What about depression? My gut would have told me depressed people respond very well to the placebo effect, but H&G say no.
I think that depressed mood may respond well to the placebo effect on a temporary basis – after all, mood seems noisy and low-bandwidth and hard to be sure of in the same way pain and nausea are. But most studies of depression use tests like the HAM-D, which measure the clinical syndrome of depression – things like sleep disturbance, appetite disturbance, and motor disturbance. These seem a lot less susceptible to subjective changes in the way the brain perceives things, so probably HAM-D based studies will show less placebo effect than just asking patients to subjectively assess their mood.
Modern authors have not disproved placebo. The problem is that most do not understand the placebo affect. The original Beecher and Ward studies are not in PubMed as they are from the 1950’s, PubMed only goes back to 1963. The original studies were a series of elegant studies which can not be reproduced today. No IRB would approve them. Consider they did sham back surgery which consisted of putting the patient under, giving them an incision, keeping them under anesthesia for 2 hours, not telling the nursing staff, and then sending them a bill. No IRB would approve. BTW sham surgery was as successful as “real” surgery. Recent studies do not actually reproduce the originals. They also showed that placebo affect wears off over a relatively short time. Double blind studies need to be against placebo to prove actual effectiveness vs the short term effect of placebo. This is known in industrial engineering studies where the act of studying improves effectiveness with no change. The placebo affect of pricing and color on pill effectiveness has been shown multiple times. BTW Cochran review is actually considered non level one evidence by many experts these days. There are multiple issues with the Cochran review, ie parachutes do not work by Cochran review procedures. No level one double blind studies only case reports. Better is classic review articles by experts in the field who know what has been published and what has not been published.
Was this back surgery to treat pain? I think that people agree that the placebo effect still works for the perception of pain.
I wonder if reading with only one eye (e.g. by closing the other eye for those of us with two working eyes) would make one more likely see the “the the”.
“Acupuncture – assuming that it doesn’t work – has been tailored over thousands of years to be as effective a pain-relieving placebo as possible.”
In other words, acupuncture does work.
Stop it with the the goddamn “the the” thing already! 🙂
I’d be careful about acupuncture. As with a lot of these mysterious Eastern things, there are a lot of charlatans about, so obviously charlatan acupuncture is going to do no better than deliberately sham acupuncture.
But there’s a confluence of modern enthusiasts’ investigation into some disparate fields (such as martial arts, deep tissue massage, yoga and acupuncture) that seems to be hovering around the idea that the body’s fascia may not be quite the inert filler material we thought, and that it might have some relevance to the immune system, as well as to other health-related topics. There’s a plausible argument that while acupuncture as a whole may be a somewhat confabulated, over-complex system overall, it could have some basis in reality, in relation to the fascia. (Could also be related to the “belly brain” idea.)
I would guess that acupuncture, self-cutting, TENS machines, and tattooing (for those who admit to using it as therapy) all work on the same endorphin premise. So, while needle placement may mean jack , I doubt the effects acupuncture are optimized placebo.
Anecdotal: As a former cutter in my youth and an adult chronic pain patient, my experience has been that controlled sensations strong enough to distract from pain feel beneficial. Using new pain/discomfort to combat old pain/discomfort. That’s why I added a very difficult exercise/sport to my weekly routine. I’m frequently in some degree discomfort and pain after the exercise–but there’s also a feeling of accompanying “high.” After a year of this addition, I’ve noticed that I’m dealing with chronic pain better. Of course, a person could argue that it’s the feeling of being “in control” of what my body is experiencing that is helping alleviate the pain sensation.
I liked Johann Hari’s essay on depression/control, though it got a lot of kickback:
My wife was literally crippled with knee pain for a while. I pushed her around in a wheel chair. Doctors diagnosed chondromalacia. She’d already been struggling with carpal pain for years that disappeared after surgery but came back with a vengeance. At some point I noticed her pain became more intense after stressful events and started thinking about psychogenic causes. I came across John Sarno’s writings. He’s a quasi mainstream figure who has cured large numbers of people from chronic pain through what is basically an openly admitted placebo effect. I knew my wife would not accept any suggestion that her suffering was “all in her head” and wondered how to broach the subject.
Funny thing… while reading forum postings from people who followed his plan I started to realize I myself might be suffering from psychogenic pain. Hip pain had forced me to quit running. As an experiment I just started running again and ignored it. After a couple of months what seemed like very real pain faded away. At this point my wife had a falling out with a close friend and her carpal predictably flared up a few days later. I showed her a John Stossel interview of Sarno and told her my experience. She immediately read one of Sarno’s books and was pretty much instantly cured of all pain. She still gets mild knee soreness but that’s it.
This experience makes me wonder how many people out there with crippling pain are in the same state my wife and I were in? I mean it’s real pain. You feel it, but it’s all in your head. I had a friend who suffered from neck pain for years. She was chronically sleep deprived. She was cured at a Pentecostalist healing revival. At the time I wrote her off as crazy. But hey she can sleep now. Read about the history of carpal tunnel. It used to be meat packing workers and such. People who did heavy gripping, not office workers. Even stenographers weren’t affected.
I thought Scott had reviewed Sarno’s book, but in fact it turned out to be a different but closely related work by Howard Schubiner. Still, may be of interest.
Scott-I believe when discussing the placebo effect it’s important to distinguish between subjective and objective improvements. Subjective symptoms, are, in my 30+ years of clinical experience responsive not just to placebos, but also to simple reassurance. I have had many patients “pain” improve quickly and dramatically upon being informed they don’t have cancer, or a kidney stone, for example.
My father drove 7-year-old-me-with-a-broken-leg in the back of his truck 3 hours from ski school to the hospital near our home thinking it was just a sprain (it had been splinted at the ski school). He told me to stop screaming and it would feel better, and amazingly it did.
He felt horrible when the x-ray showed it was a break.
Horrible for making it feel better?
Horrible for assuming it was only a sprain and discounting my pain, and horrible for driving me home 3 hours in a bouncing truck instead of going to a nearer hospital, likely (we had to cross the Puget sound to get to the hospital near home – there were a ton of Hospitals over an hour closer).
Hopefully for that reason. But I wouldn't be at all surprised if he felt horrible for telling you that stopping screaming would make it feel better, if his saying so was motivated in part by believing that it was only a sprain. Even though it helped. People are like that.
“Acupuncture – assuming that it doesn’t work – has been tailored over thousands of years to be as effective a pain-relieving placebo as possible. Maybe there’s some deep psychological reason why having needles in your skin intuitively feels like the sort of thing that should alleviate pain.”
I understand what you mean, and don’t want to be the asshole who redefines things to fit whatever point he’s making, but this wouldn’t be ‘acupuncture not working’ – it would be acupuncture operating as a legitimate treatment, whose mechanism of action involved the brain processes governing the ‘placebo response.’
This reframes things interestingly; it allows us to define some diseases as a mind-body problem rather than merely one or the other. This likely isn’t any revelation to you, but getting patients who are suffering from (say) chronic pain to understand this often makes the difference between a resentful ‘my doc said it’s all in my head, he’s ignoring my very real problem’ and a more positive ‘alright, let’s see what we can do to attack this as a more comprehensive issue, which also involves the higher nervous system and its reflexes.’
I wanted to say this too (so I guess that I wanted to be that asshole). If the complaint is that it works just as effectively if you do ‘sham acupuncture’, that is acupuncture at the wrong locations, even that doesn't mean that it doesn't work; it just means that this aspect of the traditional treatment is not required for efficacy.
I almost think that Scott was joking, given that he wrote ‘tailored over thousands of years to be as effective […] as possible’; that doesn't sound like something that doesn’t work!
This happened to me the first time I was given an antidepressant. (20mg Celexa, not that it mattered.) My brain had been stuck in a death spiral for three days, and then suddenly it wasn’t. *shrug*.
Here’s a possible explanation for asthma.
First, I don’t really know what I’m talking about, so if someone else does, I’d appreciate correcting anything I said wrong.
Anyway my adopted daughter had asthma. I noticed that it tended to come on when she was going to experience unpleasant things. For a while, I wondered if she was malingering. But I then I was reading about the disease and the literature said it is not a problem with the lungs, but with the brain. That is asthma is caused by your brain causing your lungs to malfunction.
(So I gave her the benefit of the doubt.)
Maybe the placebo effect works on asthma because it is a mental state issue.
Scott, you understand these things better than I do. What do you think about that?
Something physical (lungs) is going on with asthma though although stress can certainly make it worse. But stress makes almost everything worse. We know that it involves physical changes (lung airways narrowing and mucus build up) and some forms of it are induced by irritants to the lungs. And bronchodilators work to treat it.
Whether some nonvoluntary part of your brain or localized inflammation narrow the lungs seems relevant to treating it maybe, but not relevant from a moral perspective (“malingering”). Of course, you can also malinger while having a definitely physical illness so it doesn’t mean much either way.
I had severe asthma as a small child and had to wear a mask to deliver some sort of aerosolized drug about an hour a day. I don’t have any particuarly bad memories about it (I was maybe 5 or 6). After that, I only had to take inhalers twice a day which I did until some time in high school. I’d say one great thing about asthma as compared to other possible ailments I could have had was that I basically grew out of it; doctors told me that’s pretty common. It’s possible that if I gained a lot of weight or lived in a polluted area it would come back, but it’s otherwise not an issue.
EDIT: Incidentally, I don’t remember ever having an asthma attack triggered by something unpleasant strictly speaking unless that unpleasantness was obviously related to my breathing. Like hard exercise or severe smog. But it’s easier to end up hyperventilating with asthma I think; this may be distinct from an asthma attack. I’m not sure what the distinction is exactly.
My running theory is that the placebo effect actually used to be more effective, but that exposure to the idea weakens it. It used to be possible, even likely, that a majority of randomly-selected patients had never even heard of the placebo effect. Now that the phenomenon has been endlessly discussed in decades of popular TV shows, a lot more people are aware of what the placebo effect is. I suspect that this weakens or even prevents the effect in many cases. It’s hard to make definitive statements about one way or the other, though, since to my knowledge we still have a very limited understanding of what the placebo effect is and what causes it.
Would be happy to see a post just about regression to the mean. People talking about baseball statistics and misrepresenting regression to the mean is gonna give me an aneurysm. Regression to the mean does not mean reduced variability over time!
But if you want to be careful about definitions though: does recovery from a cold fit yours? There’s an active force (the immune system) trying, and eventually succeeding, at restoring your baseline (healthy) state; it’s not just a statistical artifact.
Discussions of regression to the mean invariably cite various fallacies, whereby people note certain patterns (actually caused by regression to the mean) and construct a causal explanation, but in fact there’s really ‘nothing’ going on after you understand regression to the mean.
But for some illnesses and recovery, there _is_ something there, something that could dominate the (present but perhaps far smaller) statistical effect of ‘sampling’ someone at the peak of illness.
A couple of decades ago I suffered from bad reflux. Bad, bad reflux. (I started to tell you the details of my experience, but realized it was too gross; suffice to say that many meals did not progress past the first bite.) I didn’t know what it was — I had experienced heartburn, but this effect was typically not associated with heartburn, so I had no clue: Throat cancer? Panic attack?
I went to my doctor and described the situation. She looked happy and said, “I can fix this!” and wrote me a prescription for prilosec (not OTC in those days). I was very busy, and it took me a couple of days to get it filled, but during that couple of days I ate completely normally, with no sign of the distress that I had been experiencing.
For a while I joked about prilosec being so fast-acting that it works two days before you swallow it.
Also, the cited paper provides further evidence that it takes us almost two decades to integrate important evidence into scientific consensus. This is depressing, but a) not surprising, and b) probably at least the sort of depressing which responds to placebos.
Maybe the placebo effect is undermined by people knowing about the placebo effect.
Does this have any carryover to the “nocebo” effect?
Nah, I’ve been using honey as a placebo for nasal congestion for a few years now. I know full well that the “honey has pollen=build up immunity!” thing is completely bunk, and yet the days I add honey to my drinks my congestion goes away afterwards where it doesn’t on days I don’t honey up.
But I also can make congestion go away by exercising, so *shrugs*.
1. Maybe honey actually does help, for some reason.
2. Maybe you believe that honey will help, meaning the placebo effect is preserved.
I was wondering about the nocebo effect – the prediction of this “no placebo, it’s regression to the mean” theory would be that it would generally apply to problems which are likely to get worse before they get better.
Can anyone say off-hand whether this might be the case?
Did she immediately stop shaking as well?
Maybe acupuncture really works but the exact points where you insert the needles don’t matter, and while the traditional underlying theory of qi flowing through meridians is bunk, creating many small wounds on the skin releases endorphins or something which relieve chronic pain.
Is there an overlap between that and the effect of getting a tattoo? I’ve heard the claim that there’s an endorphin rush with that, too.
Both tattoos and self-harm are correlated with suicidality. Maybe this endorphin rush works, to some extent, as a form of self-medication for anxiety and depression.
There’s a theory that the small disruption of the cell’s location relative to each other, or some other sort of microtrauma, might stimulate a healing effect without causing serious damage.
The endorphins theory is a decent one as well, though. I’ve had it done at some street fest thing, and it definitely gives you a good sort of ‘buzz’ in the area. Talking with a fair number of patients who have tried it, they seem to report an almost-magical cure or very little, transient effect.
This seems consistent with it having a small, positive-feeling biological effect that – in some cases – also triggers a more dramatic placebo effect.
https://sciencebasedmedicine.org/placebo-are-you-there/ is an excellent breakdown of the components of the placebo effect
> Then people started adding placebo groups, this successfully controlled for not just placebo effect but regression to the mean, and so people noticed their studies were much better
I’m sure I’m just being dense, but I don’t understand how placebo groups control for regression to the mean. Wouldn’t the participants in the placebo group be equally likely to regress to the mean as the participants in the treatment group?
Yes, and that’s exactly the point – if you have some actual drug-caused impact D, and some regression-to-the-mean-caused-impact R which is the same for both groups, then the experimental and control groups will have total impacts D+R and R, respectively. Then, to isolate D, you account for the magnitude of R, like so: (D+R) – R = D. “Control for” here just means “measure so that we can subtract it out later”.
Ah, right. Thanks.
Here’s my hypothesis: There are some conditions where it’s easy to slip into a loop where thinking about the condition exacerbates it. For example, tinnitus. If you notice that you have tinnitus, you start paying more attention to it, which makes it easier to notice it. I have a mild case of tinnitus but I haven’t actually suffered from it in years. The solution was to run a quiet fan while I sleep at night. What’s interesting is that even when I am sleeping away from home and there is no fan, I still don’t have a problem, presumably because I am not primed to notice it.
I’ve never suffered from chronic pain or clinical depression, but I wouldn’t be surprised if these are also “loop” conditions, i.e. the brain plays a significant role in the condition. So perhaps a placebo helps some people break out of the loop.
I don’t care what the studies say. I’m taking my placebo pills regularly, because the routine of taking the pills itself makes me feel better. The trick is, I take pills that I suspect are placebo only, but I can’t prove it entirely for sure. If I knew for sure, that would ruin the illusion. So homeopathy, for example, wouldn’t work for me. Vitamins and mineral supplements, both in too low dosage, are the easiest for this.
I have a friend who used vaping as a way to ease himself off of smoking– the liquid comes in various concentrations of nicotine, and he kept moving down the concentration levels. He’s now down to 0% nicotine… the funny thing is, he can’t actually quit the vaping without getting headaches, being really grumpy and tired, etc. It’s just water vapor, and he knows it’s just water vapor, but if he gives up his placebo, he gets significant withdrawl symptoms.
When I was suffering depression after being diagnosed with cancer in 1996, I found hypnosis quite helpful at improving my mood. I am, by nature, a fairly skeptical person so I hadn’t found previous optimistic pep talks about how I was going to beat this thing terribly plausible or helpful. And that’s a big problem when you have cancer because there’s a lot of work you need to do to get treated, and it’s not at all unlikely that you’ll feel instead like just pulling the covers over your head.
Fortunately, hypnosis lowers one’s resistance to influence. I wrote a specially crafted pep talk for my hypnotist to give me after she had gotten me into a more suggestible state.
At first the sessions improved my mood for a couple of hours, then for the rest of the day, then the rest of the week. After six or eight sessions, my mood problems were over.
Thanks, that was useful to me. My wife’s ex (and my stepson’s father) died of cancer likely due to that effect. I’ve always somewhat irrationally blamed him for that. I found it instructive to learn that others have struggled with the same conflict.
Oh so there is a point to all the messages we give cancer patients about how they’re strong fighters who can beat the illness. It always seemed very weird to me because i thought the patients don’t actually do anything other than let the doctors treat them. In fact, being told that i could beat an illness through sheer force of will actually seemed like it would be incredibly annoying, since it would be blatantly untrue. If in reality there is some truth to it and it takes actual effort to survive cancer, i’m going to have to move it up on my “probable cause of death” list.
With cancer, though, letting the doctors treat them may not be a trivial matter. Most cancer treatments I think involve at least adjunct chemotherapy or radiation therapy, which is basically dosing the patient with something that is marginally more poisonous to cancer cells than to all the other cells (e.g. literally mustard gas, though I think that’s mostly obsolete). This can be a sufficiently miserable experience that people may reasonably say “just let me die already”, particularly if the expected outcome is living five more years instead of two.
So surviving cancer may select for something approximating “being a strong fighter”. Not that I believe every statement to that effect is a well-considered assessment of relative treatment costs and benefits, of course.
Yes i get that now that i’ve thought about it, i just really hadn’t before. Which made me realize that if i get cancer i have lower chances of survival, because i’m the kind of person who fails the the Test of the Gom Jabbar.
It always seemed very weird to me because i thought the patients don’t actually do anything other than let the doctors treat them
Seeing a family member (who died of lung cancer), the first two doses of radiotherapy had such severe side-effects that they said “don’t give me any more, just let me die”. Doctors agreed because the radiotherapy wasn’t actually going to do anything useful (the cancer was diagnosed so late), it was more “we have to do something as a treatment and this may give a few months more”.
Dragging it out to nine miserable months of life versus six months of okay-ish if pain properly managed life was not worth it. So if you are faced with “this therapy will make you feel like death warmed up but it’s necessary to try to get the disease into remission”, the patient really does need the reserves to be able, on their sixth month of losing their hair, throwing up, feeling like they’ve been hit by a bus, and no energy, to say “No, I’m going to stick this out”.
Would you be willing to share the details/contents of the pep talk?
Placebo effect kicking in the moment a pill hits the tongue? Ha, that’s nothing. There have been several times when I’ve been of SSRIs for a while (for anxiety), I decide I need to be back on them, and making a phone call to arrange an appointment with my doctor to see about getting them seems to cause the placebo effect to kick in. So the placebo effect starts a few days before I start taking them.
When I was suffering panic attacks after being diagnosed with cancer in 1996, a lot of my panic attacks were rather meta: e.g., I’d get worried I would have a panic attack while I was on the airplane and then I’d have a panic attack over the possibility I’d have a panic attack.
Fortunately, Xanax helped quelled the panic attacks. But, interestingly, just carrying the bottle of Xanax around in my pocket soon put an end to the meta-attacks since I knew I could stop a panic attack with Xanax. After a few weeks, I never needed to take the Xanax anymore and after another month I stopped carrying the bottle around.
Maybe placebo thiotimoline would work even better.
Thiotimoline is still stuck in development, but we can sincerely promise to give all the patients a dose as soon as the pharmacy has some in stock.
Thanks, I feel better already!
Seriously the “the the” thing is so annoying (because of course I never notice it). I guess the next SSC post will have a paragraph consisting only of the word “the” repeated 100 times, and then “Ha! Failed to notice that I guess!”, and at that point I will have no other choice than throw my laptop through the window.
I’ve made peace with it. I used to think, “Wait, I’m not seeing things correctly!”
Now I think, “Okay, good. My brain is doing its job of filtering out all the stupid distractions I don’t need.”
I think your brain is going “Same short word again? Oh, I must have accidentally skipped back a word. Happens all the time. Ignore and keep reading.”
I think this is right. I’m a bit surprised Scott considers this perception rather than interpretation. Reading is basically interpreting the meaning of strings of words. The perception part of it is (I would say, as a non-expert) recognising individual words, at most.
Does anyone else find that at a line-break they’ve gone back to the same line and continued reading for a bit before realising it doesn’t make sense? I think that’s related.
> Does anyone else find that at a line-break they’ve gone back to the same line and continued reading for a bit before realising it doesn’t make sense? I think that’s related.
I did this ALL the time in elementary school — I remember getting really annoyed at myself for it sometimes. As an adult (with many years of voracious reading practice later) I still make that mistake but not very often.
I also never notice the duplicate “the’s” in Scott’s posts, but I’m not upset at this and have never put in effort to detect them until he tells me to they exist.
I’m more annoyed that seeing the title “Surfing Uncertainty” referenced yet again didn’t raise the red flag fast enough.
I hate when he does that!
Scott, did you see this?
What do you mean, this post did have a paragraph like that.
He put it on the fucking blog header now
Even that one got me the the first time I read it.
That’s OK, your comment got me the first time II read it.
Scott is an AI testing our cognitive limits. 584 days to singularity.
Can we speed it up? I’ve got some deadlines I’d rather not have to worry about.
O thou of little faith! If you believed, then you wouldn't worry anyway.
Has anyone started noticing the double ‘the’s? I haven’t
Not only didn’t I notice, but when I went back and looked for them, I only saw the first pair. I mean I looked right at the second pair and said to myself “Just a single one, he must’ve only replaced some of them in the paragraph.” I had to go back and scan the paragraph a second time to finally detect it.
Nobody ever has.
I kept missing them, until I realized Scott was obfuscating all his double ‘the’s by slipping a ‘fnord’ in the middle.
A in the middle?
Nope. I let out a huge “DAMMIT” as soon as he said I missed them AGAIN.
No. It was a “God dammit!” moment again.
There are only two types of people on this blog. People who haven’t learned to notice the “the the”s, and bots.
Not only did I miss them, somehow I glazed over that entire sentence and parentheses, so until I got to this comment I didn’t even realize Scott had called it out again.
Pretty much skipped straight from “Surfing Uncertainty” to “truck run over your foot.”
I might try writing a browser extension to highlight ‘the the’.
Now I’m wondering whether there’s something special about “the the”. Would people be as likely to overlook “a a”? “if if”?
I think good science still requires the actual placebo to be there, or at least the best comparator we can come up with. The book, Ending Medical Reversal has some good examples of cases where studies that didn’t use placebo were reversed when placebo was used later. (Granted, cases like vertebroplasty are pain cases.)
I’m honestly a little shocked that there was little or no effect of placebo, since one of the things placebo does is blind doctors as to what their patients are getting. I think if you’re going to claim placebo is mostly useless, in all but certain specific cases, wouldn’t you have to also argue that blinding is also mostly useless? But the double blind is also meant to prevent the introduction of errors due to investigator bias – which is a different mechanism. How would that square with the specific indications mentioned?
Even if placebos aren’t effective, experiments still need to use control groups. This is merely arguing that ‘no treatment’ for a control group would be the same as ‘placebo treatment’. However, I think your point about experimenter bias is important – for drug therapies, a placebo pill is the simplest way to keep experimenters blind to who is in which group.
Yes, an important use of placebo in drug trials is to blind the researchers, not just the patient.
Presumably in a study done to determine if placebo is effective, experimenter bias effects might work differently than in a study to determine if a particular drug is effective. (I haven’t looked at the designs of these studies)
In the absence of specific evidence on your point, I would definitely hesitate to do drug trials with “placebo-less” controls (and hence without blinding!)
I work in clinical research. The most popular way to conduct blinding is through a computerized system. All bottles of drug are given numbered labels and labeling is conducted by a third-party (usually a paid vendor who specializes in this kind of thing). When a subject is enrolled in the study, the computer tells the research nurse which bottle numbers to assign to the subject. Nobody, including the entire research team, knows who has drug and who has placebo. In emergency situations, we can break the blind for an individual patient while maintaining the blind for the rest of the study population. This is very rare, though – I’ve only seen it happen once, and that was in error. And once the blind is broken on a patient, they’re automatically withdrawn from the study.
That’s why it’s so strange that when these researches analyze the effect of blinded placebo trials they find no effect. The blinding mechanism is, by design, double-blind. So to say placebo doesn’t work for patients also suggests it doesn’t matter that the researchers don’t need to be blinded. But then, what about all those other studies showing researcher bias changes results? Is it that the placebo-doesn’t-matter research is just a weird outlier we haven’t explained yet? Or is there another way to resolve the conflict?
We are in agreement that this is the weird part.
My wild speculation is as follows: suppose the experimenter effect is that unblinded researchers generally make more mistakes in whatever direction makes the study “more successful”. For a typical drug trial, this means that unblinded researchers make the active arm look better than the control. But in a trial designed to disprove the placebo effect, maybe the researchers are instead biased to make the “control” arm look better and the “placebo” arm look worse? Were the researchers in the experiments Scott cites blind to what kind of research they were doing?
It was a systematic review. In other words, they looked at studies that were conducted for other reasons, but which fit the criteria they were looking for. In this case, that would be studies that didn’t just have a treatment arm and a placebo arm, but also had a no treatment arm. In some instances it can be important to have this third arm, for various reasons. In other words, it is unlikely that any of the studies conducted were aimed at the placebo effect.
And this could, perhaps, be the source of a more mundane explanation for the study’s findings. Is there something different between studies where they feel compelled to have BOTH a placebo and a no-treatment arm, versus the vast majority of studies that have a standard placebo?
I wonder about the kind of study this involves. For example, I saw lots of pain studies, some hypertension, depression, etc. But I notice there are no oncology studies. And for obvious reasons. No oncology patient would sign up for a study that included a “no-treatment” arm, and no IRB would approve it. The nature of the analysis inadvertently selected for studies that are necessarily low-risk, since no IRB would ever approve “no-treatment” for any serious condition, and no patient would agree to it, if they could even find a physician willing to offer it to them. Maybe I’m wrong, but it seems like we start by saying, “placebo effect could be there in some indications but not others” and then the study necessarily filters out large categories of indications. From there, I don’t know how we make a meaningful general interpretation of the results.
It sounds like we agree that we don’t. At best we can make specific claims about the studies analyzed.
The first paper has this to say on the subject:
> In a low-bandwidth system like pain perception, the original guess plays a pretty big role, with real perception only modulating it to a limited degree
My understand on the modern theory of pain is that mental pre-conception actually does play a pretty big role. There aren’t dedicated pain signals, in the same way that there exist dedicated temperature or pressure signals. Instead the mind has to infer in a Bayesian way whether a sensation is likely to be doing to harm, then tags those sensations as “pain”.
This seems evident in things like chronic back pain, where the underlying source is hard to observe and the Bayesian prior is especially ambiguous. Back pain sufferers don’t actually have any higher rate of degenerated discs than non-back pain sufferers. It seems like the presentation of low back pain is highly path-dependent. One day you’re doing some movement that mentally feels not so safe (like lifting a sofa), your mind receives some sensation from your spinal region, then boom from now on that sensation is tagged as “back injury”. The same sensation if originally received in the context of some non-injurious activity (like doing yoga), gets tagged as benign.
Even explaining the above scientific understanding to patients, actively serves to reduce subjective pain rating. So it seems like even very high level mental abstraction still hook into the Bayesian pain system.
Source mostly covering sources for the above:
Was the degenerated disk theory ever more than a guess?
My notion is that most back pain is a result of high muscle tension– it’s presumably a result of something happening in the brain/nervous system, not the structure of the back.
“Was the degenerated disk theory ever more than a guess?”
So far as I can tell, not really. I see patients with chronic LBP all the time and the diagnosis is extremely common, despite lack of supporting evidence. The reason seems to be that morphological disc changes are very common, and very easy to see in x-rays and so on, so they get pegged as the culprit more often than not.
I have a single subjective “data point” in favor of this:
We were once sitting tightly clustered around a table in a bar with a bunch of friends. The girl next to me was wildly gesticulating with a cigarette in her hand, which had me on a peripheral alert. Finally, as I was engaged in a conversation with someone on the other side, one of her broad gestures jabbed the cigarette tip into my forearm. I jumped up from the burning pain with a loud “ouch!” only to have her give me a weird look and calmly inform me that the cigarette was never lit…
But my brain had definitely decided that I must have gotten burned and kept an unusual ghost of the sensation around for a couple more minutes until vigorous rubbing of the spot finally convinced it that the data was faulty.
Supporting this: I used to have awful problems with the dentist. When drilling a cavity or something, I’d always be stopping him because of pain, but he’d swear that I’m so loaded up with lidocaine that I couldn’t possibly feel anything.
What I eventually figured out is that it really didn’t hurt. But I was getting all kinds of other signals that would normally be associated with pain – the sound and vibration, the smell of my tooth dust, and so on – and thus conclude that I must be in pain. When I really cleared my head and forced myself to work through it all, I found that there wasn’t any pain as such, just those clues but with a hollow core.
Subjective pain level is highly correlated with anxiety.
… A heck of a lot of cases of back-pain turn out to be a bacterial infection in the spinal fluid. We know this because certain antibiotics (ones that cross the barrier into the spinal fluid) help a very great deal.
Next step is obviously to come up with a better delivery method than months long courses of antibiotics, because those are terrible praxis re:Antibiotic resistance. Epidural antibiotics?
Isn’t the cerebrospinal fluid supposed to be sterile and separated from everything else? How do bacteria reach it in the first place, assuming no major injury? (Wikipedia says that in most people the cerebrospinal fluid is continuous with the inner ear fluid. Maybe bacteria can get in due to a ear infection?)
> A heck of a lot of cases of back-pain turn out to be a bacterial infection in the spinal fluid. We know this because certain antibiotics (ones that cross the barrier into the spinal fluid) help a very great deal.
That antibiotics help suggests that spinal fluid infection is a culprit, but it certainly doesn’t prove it. There’s plenty of alternative hypothesis. Pain is multi-faceted, subjective and highly responsive to general life circumstances. For example falling in love causes subjective ratings of pain to decrease.
Here’s a plausible alternative explanation. Start with the fact that long-term regular antibiotic consumption increases weight and food consumption. There’s pretty good evidence of this among other things A) this effect is definitely shown in livestock, B) children who take more antibiotics are more likely to become obese, and C) fecal transplants have been shown to cause weight loss.
Maybe it’s just the case that the antibiotic group ate significantly more food. Like being in love, eating above your metabolic rate is quite pleasurable at the time. (Until you pay later with obesity-related health costs.) Maybe the antibiotic group simply reports less pain because they’re literally fat and jolly.
Before accepting the spinal fluid infection hypothesis, I’d really like to see some actual direct evidence of infection itself. Otherwise it’s just to fishy. Like the other respondent said, spinal fluid’s suppose to be sterile. And why aren’t more people dying from spinal fluid infections? Why do they all stay at the low-grade chronic inflammation level? Nearly every other bacterial infection either is cleared by the immune system or eventually grows to become life threatening. This would be a very strange pathophysiology.
Alternative explanation along the “fat” theory:
Fat people with back pain largely have back pain because they are overweight, and are straining their back.
Thin people with back pain have back pain because they have insufficient cushioning – fat doesn’t make them ignore the issue, the issue is lack of fat, which is solved when they gain some weight. (This may apply to fat people dieting, as well, who may be losing the wrong fat deposits.)
That’s a neat (as in “tidy”) explanation, even if it is somewhat rigid:
Q. Why do I have back pain?
A. You’re too fat/not fat enough!
What weight is “just right” for not getting back pain? The Goldilocks Equilibrium, as it were? 🙂
I can add some color to the “too fat” end of that theory, I mostly wear jeans and when I got fat, out of some combination of vanity, laziness and denial I didn’t like to buy new jeans and when I finally did do so I didn’t like to admit I’d gone up as much as I had size-wise, so I tended to buy pants that weren’t *much* bigger than what I had had before.
Thus, I was wearing pants that used to fit me and no longer did. These too-small pants constrained back movement and (especially when sitting down) applied pressure in ways I didn’t need; I suspect this contributed to frequent back pain which did in fact go away when I lost the weight.
(caveat: sometime in recent years it seems like standard jeans material has gotten a lot stretchier than it used to be so this might no longer be as much of an issue as it once was.)
That would not show up as a positive outcome in the one year followup. The antibiotics seem to cure the backpain permanently in a lot of patients. Including way after the end of treatment
> The antibiotics seem to cure the backpain permanently in a lot of patients.
The study you linked only seems to survey patients one year later. I wouldn’t go so far to call that permanent.
Certainly that timeframe works equally as well with my gluttony hypothesis. It wouldn’t be unusual for someone to gain 20 pounds in a year. Particularly if they were injured and sedentary.
If the treated group gained an extra 20 pounds, that translates to one extra scoop of ice cream with hot fudge every single day of the year. I think a scoop of ice cream with hot fudge every day would almost certainly raise anybody’s subjective wellbeing. And along with that subjective self-assessment of pain.
I’d want to know which antibiotics show benefit, and consider that those drugs might be doing something other than killing bacteria in the CSF.
I see no reason to have priors in favor of a drug that is antibiotic having zero effect on human cells.
The study you cite discusses possible infection in the vertebral bone and/or intervertebral disc. Cerebrospinal fluid (CSF) is considered sterile and does not communicate with either of these unless the dura mater has been violated either traumatically, surgically or, rarely, by a very serious and obvious infectious process. I don’t believe the study in the article cited tested CSF at all.
One day you’re doing some movement that mentally feels not so safe (like lifting a sofa), your mind receives some sensation from your spinal region, then boom from now on that sensation is tagged as “back injury”.
What I did was Be Stupid (tried moving a heavy wardrobe without emptying it first. I didn’t think of it as “not safe”, I thought of it as “Sure I’m strong enough to move this just the little bit I need to move it, it’s too much trouble to empty it out for the sake of five minutes”). Surprisingly, I hurt myself (imagine!) The next day I was “time to get out of bed – what the hey, I can’t move!” And it took a while to get better, and I do think that I damaged something because even years later, there’s weakness and twinges when I exert myself. I don’t think it’s imagination, but I can see why the medical profession might go “Well, there’s nothing there we can physically see, so it must be in your mind”.
I think those “oh-oh, better be careful” signals from mental inference are a way your body reminds you “The last time you did stupid crap like this you hurt yourself, don’t do stupid crap again”.
Your last paragraph sounds more reasonable than the quoted bit.
I spent years doing unsafe things (I worked in a pretty labor-intensive field of construction), but had my back go out while I was wiping down a table first thing in the morning. I can’t imagine that movement was interpreted as not-so-safe.
had my back go out while I was wiping down a table first thing in the morning
I once managed to twist my ankle very badly (as in “couldn’t walk upon it”) by getting out of bed. In between hopping around on my good foot and going “ow ow ouch!”, I was “how the hell did I manage that???” 🙂
I woke up already injured a bit to my upper leg muscles. Got you beat there, D. 😛
This may be how chiropractic care works – by creating weird, sudden motions in the spine in a setting the patient is well-assured is safe and healing. The movements then get re-associated with ‘this is an OK movement.’
I’ve worked adjacent with a few chiropractors, and with a lot of patients that have tried them, and the thing I noticed is that a certain chiro will work like magic for some patients and not at all for others, and so they go to another one who works great for them but that other people complain of. I wonder if it’s to do with whether or not that particular doctor’s personality tends to put them at ease or not – or if differences in manipulation style more or less replicate the specific movement they have a pathological association with.
I wonder if the placebo effect depends on conditioning? That is, you have an anxiety attack for the first time, your doctor gives you a real pill, it works, and you then reliably can associate “this pill will make me feel better” with taking one. Then if you get a sugar pill that looks like the real one, it works just as well because the dumb suffering animal parts of the brain are going “okay we know this works and we’re not gonna die” so it calms down even before there’s any chance of the active ingredient(s) being absorbed.
Whereas if your researchers start off with giving you a sugar pill the first time you have the Purple Ab-Dabs, it does nothing, you think “well, this new medication is useless” and then it doesn’t matter if they swear blind that this is Real Medicine, it’s not going to work. Or if you are not aware, not being a medical researcher, that you are supposed to know ‘small oval peach-coloured Sugar Pill is the same in appearance as small oval peach-coloured Real Pill for this condition’, so the ‘taking a small oval peach-coloured pill makes me feel better’ conditioning does not exist and so can’t kick in?
I wonder if it’s the opposite with placebo. In nearly every study an IRB/EC is going to approve today, a patient will have to know 1) whether it’s possible they will get placebo, and 2) what the probability of them getting placebo is. We would therefore expect this to act as a ‘reset’ function for their expectations, especially given how poor the human brain can be at practical probabilities. So for example, say you’ve been getting anxiety pills, and you associate them with “this makes me feel better”. Then you join a clinical trial for anxiety, and are told there’s a 50% chance the pill will do nothing for you. You might mistakenly adjust your expectation to “I expect 50% of the benefit from this pill” or you might more accurately assume, “I have half a chance this won’t work for me. That’s not great odds, so I won’t expect much. And besides this is experimental, so maybe it won’t work anyway.” Maybe that last consideration is blunted in a phase III study, where some efficacy data is available, but I would expect the other concerns to reduce the baseline expectation of the patient.
In other words, shouldn’t predictive processing hypothesize that the presence (and knowledge of the possibility of getting) placebo reduce the predictive response? There’s a simple way to test for this: Conduct two pain studies and measure pain soon after patients get the drug (soon enough that pK data suggests it should not yet have an effect). One study will have placebo, and patients will be told of the placebo possibility. In the other, there’s no placebo, and patients know they’re getting real pain pills. If this hypothesis is correct, you would expect blunted predictive processing if there’s a chance of getting placebo, so patients won’t report feeling pain relief until the drug actually hits their blood stream.
Yeah, I think if there’s a chance you know or assume you could get a sugar pill, that will not work for the placebo effect. It all depends on the belief that “this is real medicine, and real medicine makes you feel better, and real medicine takes a while to work”.
If your expectation is “huh, some crummy junk new pill that I bet isn’t any good”, then I would certainly expect a report of “this did nothing for me” even if it’s real new drug and did have some slight effect.
I don’t think it’s entirely like that, but if you’re doing a placebo-controlled clinical trial you have to remind patients, “this medicine is still experimental” which is going to impact their biases, and therefore should theoretically also impact their predictive processing.
I wouldn’t go so far as to characterize it as “this is probably crummy anyway, but they pay me $50/visit so I guess I’ll take it”, but rather, “hope this helps my pain since I don’t like my other options; but the study nurse reminded me this is experimental and I could get placebo; maybe it won’t work as well as that commercial I saw with people walking through fields of sunflowers.”
I believe they actually have done something like this – exposing people to the object of their phobia while on anxiolytic medication. It seemed to have quite promising results, though I haven’t heard of a follow-up study yet.
I used to take a medication that, if I forgot to take for a day or two would give me minor withdrawal effects. When I next took a pill, I’d almost invariably feel some relief immediately after taking it, well before any chemical effect could actually take place. So yes, definitely, at least for me.
Even if you’ve never had anxiety medication before, you’ve probably had the experience of taking a pill that relieved some other condition.