How powerful is the placebo effect?

[drwex]

For quite some time it's been well understood that the so-called "placebo effect" (*) can have statistically significant effects on self-reported metrics such as peoples' level of sadness/happiness, confidence, and so on. Roughly speaking, the placebo effect is about as powerful as antidepressant pills in all but the most depressed people, leading to a question of whether the pills are in fact doing anything at all.

But scientists have been slow to accept (and there hasn't been much evidence for) the placebo effect making a difference in objectively measured criteria. Like, say, the size of your hips. This morning an NPR story (http://www.npr.org/templates/story/story.php?storyId=17792517) pointed me to some research published last month (NY Times story here: http://www.nytimes.com/2007/12/09/magazine/09mindfulexercise.html?ref=magazine) by Harvard psychologist Ellen Langer that may push this idea further into the mainstream.

In this study, Langer and a student talked with hotel maids about their level of exercise. Despite being in jobs that kept them physically active all day and often involved quite demanding manual labor these maids didn't see themselves as "exercising." Two-thirds said they "didn't exercise."

Langer then measured externally visible metrics of physical health: BP, waist-to-hip ratios, etc. Then they educated about half the maids on the amount of exercise they were getting simply by performing their normal daily jobs. The other half got no education.

One month later Langer re-measured the same physical metrics and found that for at least three of the metrics there was a statistically significant change. The women who were told they were exercising showed the kids of physical improvements you'd expect from a comprehensive exercise program.

Unfortunately, the study doesn't appear to be air-tight. It's quite possible that the educated group of maids did things differently; perhaps they were more mindful about things like their diet, etc. The placebo effect might not be the thing responsible for the observed changes. But it might. Even if it doesn't account for all of the change, it may well account for some. Fascinating.

Neither story tells me where the research was published, but it did lead me to discover that Professor Langer has published a couple of pop-science books on the topic, and I'm ordering those. I'll write more once I've read them.

(*) The term is accurate but some people view it as a pejorative in the sense of "it's all in your head - you're making it up." Really that's not what's meant. The effect is real and measurable. I'm not intending anything pejorative here.

Comments

[halleyscomet.livejournal.com]

I wonder how this would impact the perceived effectiveness of prayer. Just believing a painkiller will relieve pain can cause the brain to release natural pain killers (http://health.dailynewscentral.com/content/view/0001546/35/). If just believing you're getting exercise can change your body, then perhaps just believing someone is praying for you can, in and of itself, cause your body to rally and make a recovery that would have not happened otherwise.

Of course such a suggestion will get all sorts of people riled up, in part because the last I checked there was no scientific evidence suggesting prayer by a third party aids recovery.

[drwex]

I'd have to go dig up the references, but yes this has been studied. The conclusions were as you'd expect: if you know (or are told) people are praying for you then there is a marked shift in attitude and a measurable though less marked shift in outcomes.

If prayer is done but the person being prayed for is not told then there is no measurable change in attitude or outcomes.

Interestingly the shift happens regardless of whether the people being prayed for say they believe in the power of prayer or not. This is probably related to another effect (named for the scientist who first discovered it in studies of telephone operators and whose name I'm also forgetting) that people do better when you pay attention to them, regardless of any other factor you can find.

[halleyscomet.livejournal.com]

It sounds like as human beings we tend to do better when we know someone cares. It makes sense given our tendency to be social creatures.

[drwex]

It doesn't even have to be care - it's the being-paid-attention-to.

[chienne-folle.livejournal.com]

One thing I find interesting is that a lot of different animals seem to thrive when given human attention. I can see where HUMANS would have evolved to do better with human attention, but parrots? elephants? other non-domesticated animals? Why do they enjoy human attention?

[agentotter.livejournal.com]

Well, I think there you're getting into a whole other issue of what the animal's other options are. Your for-instance species, in particular, are very social animals, especially animals like elephants whose whole existence is so tied into family bonds, and who by virtue of their circumstances have turned to human affection as the only affection available. (Either that or we aren't separating the animal's perception of the human's attention from the human's perception of the animal's reaction. For instance: say I have a horse named Sparky. I go out to the paddock to visit Sparky, and he comes running to meet me and is clearly excited to see me. I think that this means he likes me, but I'm utterly failing to notice that Sparky's approach and posture indicate that his demand for affection is a display of dominance, and my giving him what he wants is my indication of submission. So is Sparky thriving on human attention, or has Sparky come running to see me because he's frustrated by his lack of exercise and disinterested pasture-mates, and he's overjoyed to see me because I'm giving him the opportunity to be dominant over someone?)

And I think it's also worth noting that the animals you mentioned specifically, elephants and parrots, are also two of the ones who suffer most obviously from mental disorders in captivity, from obsessive and compulsive behaviors to self-destructive activities and violence against people. So are they receptive to human attention because they want it, or are they seeking it out out of frustration stemming from being out of their natural environment, away from their social groups, etc.? You're asking an interesting question but it seems like there's a lot of assumptions inherent in it... before exploring for an answer I'd love to see some evidence that any non-domesticated animal (or any animal at all, for that matter) as a species derives specific benefit from human attention/affection, and not just attention in general from any other being at all. Have you seen any studies on the subject? I'd be really interested to read them.

On a sort of similar topic, there's an excellent book about how plants have evolved to be pleasing to human beings... it's called The Botany of Desire. You should check it out if you haven't read it, I think you'd find it really interesting. It centers on the concept that while we think we've selectively bred certain plant species to be pleasing to us, by the same token the plants have been evolving to make themselves attractive to us, the same way they make themselves attractive to bees, in order to more effectively reproduce themselves... so who's using who? It's an interesting topic and a really enjoyable book.

[keyne.livejournal.com]

This is probably related to another effect (named for the scientist who first discovered it in studies of telephone operators and whose name I'm also forgetting) that people do better when you pay attention to them, regardless of any other factor you can find.

Hawthorne effect.

[drwex]

You win. http://en.wikipedia.org/wiki/Hawthorne_effect

Thank you.

[sykotropic.livejournal.com]

My personal experience completely agrees with the study. My 3.5 month old 20+lb is schlepped around everyday at minimum one mile usually closer to 2-3 miles. I have two main ways of toting him. One is in a backpack and one is in a stroller. If I mindfully use my errands as an exercise time then I put him in the back pack and keep a decent pace for the walk. If I'm not mindful or need the carry space of the stroller then I use the stroller. The difference in my caloric output (measured by muscle fatigue and heart rate) is significant. So, whether its the placebo effect of telling myself I'm exercising or whether I'm just being more mindful of using the time and activity more efficiently I don't know any more than Langer does, but either way it does work.

[drwex]

I hear you. I wonder if the maids have a similar choice - I imagine their carts weigh the same and must be pushed the same distance each day, no?

[sykotropic.livejournal.com]

Which brings to mind the image cart races down long hallways.

[rednikki.livejournal.com]

Thanks - this is really useful!

OK if I link?

[drwex]

Link away. In general you're free to link anything I didn't friends-lock.

[chienne-folle.livejournal.com]

Fascinating. Thanks for passing this along.

Ellen Langer is a well-known name in Psychology, so this study was even done by an especially trustworthy person. Interesting.

The thing that makes me think that antidepressants actually work (aside from my own experience) is that people often don't get better on the first drug that's tried but have to try two or three or ten different drugs. If there was a placebo effect, why would it only kick in on the fourth drug, instead of on the first one?

[lyonesse.livejournal.com]

placebogenesis has been found to be about 1/3 of the pain-relief effect of hospital-given analgesics (studied as consumption relative to other medication made available to patients). and there's mri showing where it's happening.

not to say that it isn't a second-order rather than direct effect (which is what i guess you mean by "not air-tight") but then, who cares how much about that, relative to a clinical benefit?

i'd suggest searching pubmed for your prof langer though rather than reading pop science, though. the lamest stuff ever gets out as pop science :/

[drwex]

What I mean is that the current study is not air-tight as there were (according to the critics) insufficient controls on other confounding factors; for example, dietary change was not monitored.

[lyonesse.livejournal.com]

then the critics are abusing the terminology; diet was not an experimental factor at all -- only education was. if diet turns out to be the mechanism by which the experimental factor works, that's charming, but it's not a sign of any lack of "air-tightness" in the experimental design.

[drwex]

Err, no. That's not the point. If diet is the factor that explains the change then the study is no longer evidence for the placebo effect causing changes in observably measurable metrics. That's the key question here.

As I understand it (and I take your point that one ought to read the real published research, if there is any) the experiment showed a change in a population and attributed that change to one factor: education and change of viewpoint. If in fact the change was due to something else, that's a confounded experiment design and requires the experiment to be rerun with tighter controls (which I'd support anyway on the theory that reproduced results are better than a one-off).

[lyonesse.livejournal.com]

attributing the change to a manipulated variable (that is, a factor) is actually a-ok.

of course, behind every factor there is a mechanism (incl. in placebogenesis, which we are finally beginning to understand via mri). if educated people attend more to their diets, then that's a mechanism, but it does not change the education's status as a factor.

and of course all science has to be reproducible. the standard p-value is >.05 cutoff, which is to say that one in twenty publishable results is purely coincidental.