Great piece, Rob. I've witnessed the placebo effect in others over the years and reached simlar conclusions. I know one woman who went to her doctor's office every week for "treatment." When she arrived there, they simply did a procedure that was entirely a placebo to manage her as a difficult patient, yet every week she acted like they had done something therapeutic. Traditional medical practitioners do oft times learn to use placebos to great effect.

Thanks for this thoughtful piece.

Rob, thank you for this fascinating post! In the past I had mostly heard of placebos in connection with drug trials. Interesting to read far more about it, along with CAM therapies and the Hawthorne effect.

Thanks, Kathy. Bausell writes that placebos were actually in common use in medical practice up until the middle of the 20th century, when a refined sense of medical ethics changed what doctors do. I've also seen in person therapies suggested by doctors who would say, "There's no strong medical evidence that this works, but some people find it effective and no harmful side effects are known." That's fine, I think, respecting their responsibility to the patient.

Hi, John. I find all of this stuff fascinating. Even the details of how to design an appropriate study of a placebo-controlled therapy. It turns out that you need a control group and two treatment groups for a good test. The control group gets no treatment, one treatment group gets the therapy under consideration, and the other treatment group gets a sham therapy that is as close as possible to indistinguishable from the real thing. It's the only way to prevent the placebo effect from being mistaken from a real therapeutic effect. Even then, there are potential problems with sample size, patients dropping out, experimenter bias, and so forth. Very tricky.

Great piece, Rob.

I'm pretty passionate on this subject. I work with lots of physicians, and I understand full well that there are all sorts of problems with modern medicine.

But I also know that people are easily bamboozled, and when they're sick or in pain, they'll try anything to get better. I have nothing but disdain for the purveyors of alternative medicines who prey on weakness and uncertainty and fear, and think homeopathy is perhaps the worst of the worst.

People who push alternative medicines love to talk about all the plants that have been used as remedies for centuries, and accuse the drug companies of trying to suppress natural cures. That is, of course, horseshit. If herbal cures worked, the drug companies would be the first to commercialize them; in fact, they have commercialized them successfully whenever possible.

In other words, there's a name for herbal cures that actually work. They're called "medicine."

Thanks Rob, for the data behind the CAM story. You always write such cool stuff.

Hi, Frank! Thanks for bringing up a phrase I was trying and failing to remember: In other words, there's a name for herbal cures that actually work. They're called "medicine."

You've probably run into a lot of intelligent people, both patients and doctors, who honestly believe in these therapies. I have, too. I think it's the bamboozlers who spoil things, in part, as well as the history of our misunderstanding and even discounting placebos. For example, it would be nice to say, "This therapy is an effective placebo for some people," without the implications being "Oh, your problems are just in your head," and "You're so gullible." But these implications seem to be inevitable right now, and they make honest people feel defensive.

Thanks, OEsheepdog!

Interesting post Rob. Don’t you wish you had thought of the Hawthorne effect experiment? That has got to be one of the most fundamentally important phenomena ever delineated in research on humans.

What many savvy mainstream medical practitioners do these days is ask patients to keep them informed of any “alternative” treatments they are using rather than minimizing or denouncing their value to the patient. That way they can observe and monitor possible interactions and therapeutic effects or problems. I guess it’s the equivalent of “If it’s not doing any harm and the patient says it works, then let’s see.”

I love that final section on selecting a placebo that works. Of course, I love paradoxical events too. And I saw a survey the other day that said “on average, half of the respondents rated the product lower than the other half.”

Good Sunday morning read. Thanks.

Hi, grif!

Don’t you wish you had thought of the Hawthorne effect experiment?

Oh yeah. I mention this to my students when we go over experiment design for evaluating user interfaces, warning them that if they're not careful, they may see differences that aren't real.

What many savvy mainstream medical practitioners do these days is ask patients to keep them informed of any “alternative” treatments they are using rather than minimizing or denouncing their value to the patient. That way they can observe and monitor possible interactions and therapeutic effects or problems.

This is very good, I think. You've also reminded me of something else Bausell mentions, which is that another reason it's hard to evaluate CAM therapies is that they may be integrated into more conventional medical practices or lifestyle advice--if a CAM therapy involves X, Y, and Z, plus avoiding fatty foods and taking some time out each day for meditation, as an integrated set of recommendations, it may be hard to tease out which is effective and which is not.

Hi, Aaron!

It sounded very interesting, fascinating actually, but seemed to be a significantly methodologically challenging.

I agree that there are huge difficulties in evaluating this sort of thing, and that your point about individual differences is also well taken. There's so much variability in experiments that involve human beings as subjects, and often so much inference required to get at real causal explanations, that interpreting results can be hard. To the point that some philosophers dismiss some of the social sciences as being sciences at all--fortunately for me, the areas of cognitive science that my work touches on generally go with measures that are fairly consistent and reproducible across a wide range (like reaction times to stimuli).

Yeah, you hit all the points with alternative meds. It's a hard topic, and that last line is perfect.

Fascinating and well written. One thought I've had about pain in particular, is that I feel better when I "do" something about it, so the 'placebo effect', in that case is embedded in self efficacy more than a belief in an external effect. So, I may feel better when I take a walk, if my back hurts, I also feel better if I take medication, but in either case, the agony of helplessness is somewhat relieved by taking action.

Rob, I'm afraid you've lost me here. Just because "it works for me" is not enough evidence for the scientists does not mean that it should be dismissed. The fact that scientists refuse to allow for other means of healing just shows that their rather narrow methods don't explain everything about the body. It seems rather obvious to the people for whom alternative therapies have worked - and worked consistently and well - that it's the brand of science being applied in the research that has limitations, not the other way around.

Thanks, Julie, I'm glad I didn't get anything wrong, as far as I know; a nurse is much more likely to be on top of these sorts of issues.

Hi, Risa. That's interesting! I find the same thing, come to think of it. I've read that the feeling of control over one's environment can have a significant effect on one's health, both mental and physical. It's the latter that might surprise some people (like me, sometimes), but there's a lot we don't know yet about the connection between the mind and the body.

Hi, Ardee. You raise some interesting points that I think have to do with perspective, and I'll try to clarify my view.

Just because "it works for me" is not enough evidence for the scientists does not mean that it should be dismissed. The fact that scientists refuse to allow for other means of healing just shows that their rather narrow methods don't explain everything about the body.

I'm not saying that "It works for me" should be dismissed. I am saying that, at least according to what I've read in Bausell and elsewhere, that (a) CAM therapies are indistinguishable from placebos and (b) doctors should be careful about prescribing them, if that's what they are. For an individual, what works works. Scientist are interested in why that something works, and for the kinds of CAM therapies I've been reading about, the "why" is "they work because people believe in them" rather than because of some inherent property of the therapies themselves. That may not matter to an individual, for whom getting better is the main thing, but it matters to doctors and scientists.

Why? I think it's because we (some of us, at least) want our medical treatment to be based on our best understanding of cause and effect. If a CAM therapy is ineffective, and it's really the patient's belief that drives recovery, that's something doctors need to know, or they might find themselves in the position of recommending a treatment that's irrelevant to health rather than something that works independently of the patient's beliefs. In other words, I don't want a doctor to have to read my mind or judge the strengths of my beliefs to make a recommendation about a treatment. He or she should just tell me what works.

That's not to say that CAM therapies are useless: millions of people think they're worthwhile. It's just that, to the best of our knowledge, CAM therapies generally don't work the way most of those millions of people think they do, via some verifiable physical effect. Instead, they work via a placebo effect.

I'm finding as much food for thought in the comment stream as in the original post. "Placebos' are no longer prescribed in mainstream medicine - it's considered an unacceptable practice. As stated, it's 'sham medicine'.

This is why I was drawn into Grif's summation statement. "What many savvy mainstream medical practitioners do these days is ask patients to keep them informed of any “alternative” treatments they are using rather than minimizing or denouncing their value to the patient. That way they can observe and monitor possible interactions and therapeutic effects or problems. I guess it’s the equivalent of “If it’s not doing any harm and the patient says it works, then let’s see.”

This was said in different words by Firestorm and Rob StA and I am in agreement with encouraging patients to do what they feel is working (as long as it's not harming) and also to combine more traditional medical approaches to 'work with' the possible placebo practices.

You raise some good points, Aaron, methodological and philosophical. Here are a few narrow responses:

I agree that proxy measurements are useful to the extent that give insight into physical mechanisms. Some proxies have a lot more noise than others, and while we have statistical tools that can account for noise, I think that sometimes we may reach the point of saying, "But are our findings in any way useful?"

I have a similar memory of the cholesterol drug you mention, though I can't recall the details. This is a methodological issue for drug trials, getting at individual differences, and I think that it can be handled in principle, even if it would be very, very expensive to do so. So maybe not in practice.

I'm reminded of another point about individual differences. An analogy may be helpful: Political maps of the U.S. are often broken down into red and blue, conservative and liberal states. It's a mistake to assume that everyone in a red state is conservative or everyone in a blue state is liberal, of course. And yet people do this without thinking about it. Similarly, if we have proxy measures that are very far removed from physical mechanisms, we may be making a comparable mistake when we generalize some finding over a large number of people--it may apply to some but not others; the same result may be produced by very different mechanisms in different people; and so forth. I don't know how to account for these sorts of things either. I think it's inherent in the kinds of abstractions we're willing to deal with.

Thanks for adding your perspective, Gabby, which sounds very well-informed.

I'm not sure anyone has ventured into explaining what an experimental design looks like and why it either validates the intervention or not. Let me try.

There is an intervention (which could be anything -- a new medication, a device or a therapy) that you are testing to see if it improves pain. You take a large enough group of people that meet certain criteria that need the intervention and you randomly assign (think flip of the coin) them to the treatment or to the control (non treatment) group. Because of the magic of statistics (numbers), if the group is large enough, the two groups of people will be essentially the same.

The people don't know which group they're in. The researchers don't know which group they're in. The clinician doesn't know which group they're in. You apply the sham OR the real treatment and measure the outcomes (was pain reduced or not).

I have seen devices or medications that people would bet their life on *not work* over multiple experiments, even though there were variations in the outcome measures or the treatment protocols. What can happen is that the treatment and sham group both experience great outcomes, however when the two outcomes are compared, there is no real difference.

Lots of PhD, MD and other researchers grumble about Blinded Random/Control Group Designs (RCTs) and say that they're not necessary. However, from what I've seen, they're absolutely essential so that we can really tell what works and what doesn't. Talk about reducing health care costs.

denese

Placebo means that the patient has a positive (read: pain or anxiety reducing) response to any given treatment regardless of an objective measure confirming that response. Nocebo means that the patient experiences a negative response instead (few people ever talk about that one). The part that doesn't get addressed often enough is that pain and anxiety and fear are subjective and no one has figured out an exact biochemical marker that correlates with it. Valium doesn't fix anxiety, it just temporarily drugs your brain out of feeling it. One could say that meditation is a placebo, although many studies have been done showing its impact on reducing blood pressure, increasing lung, brain, and heart perfusion, and change in mood stability. (If you can't find the studies doesn't mean they aren't done by reputable people). The thoughts that help create and sustain a chronic illness pattern (some individuals are more prone to fibromyalgia and RSD than others) need to be changed. They induce the proliferation of kinases and cytokines that activate the inflammation and apoptosis pathways that are associated with chronic disease and illness. That doesn't mean they caused them, but pain creates a "new" disease over time with mental and physical components. New, because it needs more than a pain killer because it has changed its biochemical morphology in your body.

Unfortunately, when a natural therapy "fails" to "cure" a patient they assume they have been duped. Cancer, for one, is a name that is given to encompass many different genetic and biochemical causes that result in uncontrolled cell division. It is not one disease, or one cause.
Drug studies are only done on patentable medicine, and that excludes almost all natural substances. Patentable means that it will be exclusively owned and profited on by the shareholder. The potential profit pays for them, they are very very expensive. And as we have learned via Vioxx, many of them are skewed by the researchers.

Thanks for your explanation, Denese. I couldn't have done it so concisely. (That is, I tried but then crossed it all out. :-)

You raise some good points, Oryoki Bowl. To the extent that I know about what you've discussed, it all sounds right to me.

I've always felt there were two dimensions to healing: the emotional dimension, and the physical dimension. Science focuses on verifiable physical dimensions, and disburdens questions about the emotional dimension into the "placebo effect" bin. My sense is that someone needs to come up with a tool that can examine this emotional dimension - with a goal to increase - not sabotage - its effectiveness. To date, Science doesn't seem to have developed the relevant tools.

I've always been fascinating by the notion of placebos. But about six years ago I heard and read (both, but can't remember the two sources anymore, despite great effort to recall) that in researching the history of the Placebo Effect, some doctor discovered there really was no such thing. From what I recall, he went back to discover the science behind the thing and discovered that there was always this assumption that the Placebo was a real effect but there was never any actual controlled experiment to establish it, despite popular opinion to the contrary. Have you heard this, Rob? Does this author mention anything like this?

Re the Hawthorne Effect: For some reason, this makes me think of the situation where one drives behind a slow driver, goes to pass him on the left, and suddenly he speeds up, making your passing very difficult! It's like he knows he's being studied.

Re your explanation for a placebo effect in evaluating pain medicine for an inconsistently painful condition: I think this is the core of many, many unscientific conclusions, including attributing positive things to prayer. Most conditions or bad things pass/get overcome with time, and praying or whatever anyone does will "result" in a passing of condition. But that's only b/c time has passed. One thing I'll add though: your hypothetical relates to pain; what about placebos that supposedly work for something less subjective? Aren't there examples where actual healing takes place--let's say the fusing of bone or the elimination of cancer cells--when someone takes a placebo? This is more interesting to me. I don't have real information about it, but I thought there was some of this out there.

Finally, re what's called a CAM: I think it takes time for something once called CAM to be considered traditional. But some things seem to stubbornly remain in the CAM category even if there's evidence of effectiveness. Most of us would think that magnet therapy was a CAM, but I remember reading about two specific studies that showed that magnet therapy was as effective for carpal tunnel syndrome as surgery. So why do we still consider magnets CAM? I'm wondering if it has something to do with people then wanting to extrapolate a therapy into unproven territory. In other words, it's a CAM here but not there.

anything that works only because people believe in it would seem to me to be evidence that medical science and pharmaceutical science in particular are a very long way from understanding the fundamental mechanisms of healing on the one hand, and the power of faith on the other

good post

Rob, I'm not an herbologist, or a practitioner, nor am I researcher, so I'll not try to reach beyond my knowledge. I'd like to address this statement: It's just that, to the best of our knowledge, CAM therapies generally don't work the way most of those millions of people think they do, via some verifiable physical effect. Instead, they work via a placebo effect.
Placebo is a scientific construct, not a true result of research. From my understanding there are conflicting views even in the scientific community about what the placebo effect really is. (Personally, I think scientists toss everything they don't understand into the placebo effect, but like I said, I don't have the data to prove this) The concept of the way most of those millions of people think they do, via some verifiable physical effect though, I can say something about. Many drugs come from herbal sources, and they come from drug companies who did testing on them to find out what the specific molecule in the herbal source produced the effect. My brother-in-law works for Big Pharma, and he told me that if they can't isolate that molecule, they can't patent it, therefore they can't pass it through the FDA process. It's partly ownership but it's also the idea that there is one active ingredient that can be identified and then mass-produced. The problem with studies on, say, herbal remedies, is that they often work in synergy with other herbal substances, and different people respond to different substances. That makes it difficult to isolate, patent and develop a reproducible production process, what big business depends on. Pharmaceuticals have many side effects, because they are prescribed to everyone with particular symptoms, but not everyone has exactly the same kind of problem, metabolism, or lifestyle - one pill does not fit all! Medicine, and doctors, depend on the idea of the silver bullet, but they ignore the possibility of a more subtle regimen being more effective for an individual. It just doesn't fit into the current scientific mindset - therefore, it must be placebo.

the main problem with herbal remedies in the USA is that they are not regulated with the same stringency that we have for "normal" medications. They have, in the past, been inconsistent both in results and in ingredients from batch to batch. This does not even look at the differences of the drug between companies. If you buy generic citalopram, you get the drug citalopram. It will have the same efficacy from one drug co to another and the same chemicals in it- it's not the same with st john's wort. That doesn't mean st john's wort doesn't work. I think they've proven in Germany it does, but it's not regulated here, so you take your chances.
http://nccam.nih.gov/health/supplements/wiseuse.htm
(not a medical opinion, but a personal one)

Thanks for the comment. Internet connection is down; back later.

Very interesting indeed. I am always interested in learning about alternatives to healing. It may be a "placebo" effect when healing occurs from these non-traditional methods but what I am bothered by in these studies is that any results are summarily dismissed by putting them under the label of "placebo". I believe the mind is a most wonderous and powerful thing and while a therapy may be a placebo, it was still the catalyst for the mind that brought about the healing. That has to be worth something.

I don't do any alternative anything. I take prescribed medication for ailments including chronic pain. If I took a placebo and it stopped my pain, even if it's in my mind, why would I quit taking it, unless it quit working. Who cares what works as long as it does work and does no harm!

I'm back!

Thanks for commenting, M. Chariot and Roy. I think that the points that you raise are similar, and worthwhile. We don't know nearly as much about the healing process with respect to our minds, in particular our emotions, as we do about our bodies. We (well, not me, but people) are working on it, though. Some researchers, like Antonio D'Amasio, have even written books about their work accessible to lay readers.

Hi, Lainey.

Bausell does spend two chapters on the placebo effect, one outlining indirect historical evidence for its existence, another on direct experimentation (some of it biochemical) on the phenomenon. I don't have the background to independently evaluate his conclusions, of course, but he writes that the placebo effect definitely exists. Exactly how it works and the size of the effect, though, are still open research questions. (One interesting thing Bausell brings up is that two placebos for the same condition may differ in their effectiveness based on some obviously non-therapeutic feature--the color of the pills, for example. A placebo effect accounts for this nicely.) I imagine that there may be respectable opinions on both sides of the matter, though.

On the "core of unscientific conclusions", Bausell starts off with a discussion of cognitive biases in the context of evaluating anecdotal medical evidence. I agree that human fallibility in assessing causal relationships is exactly the difficulty. As to whether the body heals itself in ways that current medical knowledge can't account for, I've heard lots of stories about this, too. I'd guess that it happens.

Finally, on magnetic therapy specifically, we'd have to look at the study. Bausell finds that most studies of CAM therapies lack proper experimental controls, draw unwarranted conclusions, and are in general not very trustworthy. Further, because the convention in many areas of statistical evaluation is to consider a result "positive" if there's only a 5% chance that the result could have arisen from chance, and because of publication bias (negative results not getting published), false positives are a problem.

Thanks for expanding on your earlier comment, Ardee. I think I get it this time around. I hope I didn't sound too dogmatic earlier. As I understand it, the argument is that a given CAM treatment may work in exactly the physical or biochemical sense that we'd expect from a conventional treatment, but because of the limits of our scientific knowledge about physics, biochemistry, and so forth, and due to individual differences, we can't demonstrate that the CAM therapy works broadly enough for conventional experiments to detect. I find this argument sound; it's possible that this is the case.

Still, I think it's reasonable to consider other explanations. (Skepticism is built into research.) If there is such a thing as a placebo effect, as I mentioned above, is it a better explanation for the efficacy of CAM than the argument above? And is there any way to know that the argument above holds for the case of a specific CAM treatment and a specific individual? There's been some science news in recent years about plausible conclusions being drawn from individual experiences, under careful study, so I know that it's possible; I think it's also controversial. But the upshot is that the use of CAM therapies isn't necessarily unscientific or irrational; it may be a judgment call whether there's enough information to evaluate them properly.

I agree, Painting the Stars, that if we could get rid of the negative connotations of placebos, that would be worthwhile. I mean, "Oh, it's just a placebo" may be missing the point--isn't it cool that believing something can have a positive effect? I'll mention the caveat, though, that in a lot of placebo studies, relief is temporary and in fact not significantly different from having received no treatment at all. That's one of the big dangers, if a placebo is not used because nothing else works, but if a placebo is used instead of something that does work.

Scanner, I think that's a reasonable attitude: if something is not harmful, and at a minimum it increases peace of mind, it can be worthwhile, even if it may not be a purely rational thing to do. People aren't purely rational.

Thanks again for sharing your real-world knowledge, Julie.

Damn CAM. (but yay TAM!)

Energy is potential work, not ghostbusters plasma. I sometimes think this magical thinking error is at the heart of all CAM pseudo-science. If we start our kids right, make an entertainment of Logical Fallacies, get them to exercise their critical thinking skills starting in Kindergarten, the power of scammers will evaporate in a generation.

Oh wait, if we teach critical thinking at a young age they will wonder about the sky pixie too. Dang.

Fundies and New age charlatans, the unholy alliance of our age.

Excellent exegesis, Rob. A great, efficient trip through this material. I expect to see your work in Skeptic and Skeptical Inquirer some day! It's as good-better than most of what they publish. Clear, genial, un-polemical.

Thanks, Greg. I count your comment as high praise, which I'm very happy to read. I haven't been on OS that much lately, but sometimes I come across and idea that I just have to write down...

I'm fascinated by this subject, but also not sure about what constitutes a CAM and what doesn't. When I sense I may be getting a urinary tract infection, I drink cranberry juice. It seems to help, and I'm under the impression that there is plenty of evidence that it does. Is drinking cranberry juice CAM because I am not taking a pill? What about taking acidophilus for yeast imbalances/irritable bowel etc.? Is it CAM because I buy it in the health food store and not the pharmacy?

Hi, Siobhan. There's a national center funded by the NIH on CAM, and here's what they say:

CAM is a group of diverse medical and health care systems, practices, and products that are not generally considered to be part of conventional medicine. While scientific evidence exists regarding some CAM therapies, for most there are key questions that are yet to be answered through well-designed scientific studies—questions such as whether these therapies are safe and whether they work for the purposes for which they are used.

Bausell concentrates on things like acupuncture, chiropractic, homeopathy, hypnosis, and such. I think he picks these because they're in common use despite the lack of experimental evidence demonstrating that they work the way that they're said to.