Langer and Crum – Exercise and the Placebo Effect. Or: Advice improves physical measures of health? WTF?

October 3, 2008 at 8:38 pm (Bad Science, Placebo) (Exercise, Placebo, Psychology)

Ben Goldacre wrote about a study performed by Alia Crum and Ellen Langer in his book Bad Science. This study also featured on the Bad Science blog, and the story goes something like this:

Alia Crum and Ellen Langer from Harvard psychology department took 84 female hotel attendants in 7 hotels. They were cleaning an average of 15 rooms a day, each requiring half an hour of walking, bending, pushing, lifting, and carrying. These women were clearly getting a lot of good exercise, but they didn’t believe it: 66.6% of them reported not exercising regularly, and 36.8% said they didn’t get any exercise at all.

Weirdly, measures of their health (BMI, blood pressure) were related to how much exercise they thought they took rather than how active they actually were. The researchers divided the attendants into two groups and one got an hour-long presentation on how much exercise they were really getting (including the information that they were satisfying the surgeon-general’s recommendations for exercise and examples of how their work was exercise), while the control group was given no advice on their exercise levels. When, after a period of four weeks, the researchers looked again at the measures of health they had examined at the beginning of the study, they found that each measure had improved in the group that had been given the presentation while the group that had been left in the dark had no change in their measures of health. Here’s how the authors put it in their abstract:

Although actual behavior did not change, 4 weeks after the intervention, the informed group perceived themselves to be getting significantly more exercise than before. As a result, compared with the control group, they showed a decrease in weight, blood pressure, body fat, waist-to-hip ratio, and body mass index. These results support the hypothesis that exercise affects health in part or in whole via the placebo effect.

In his book, Goldacre writes that the findings of this study are “bizarre” and suggests that if anyone
has a good explanation then they might like to blog it or write a letter to the journal. I don’t have a good explanation, but I have been thinking about the strange finding that advice itself rather than changes in routine can affect these physical measures of health (weight, blood pressure, body fat, waist-to-hip ratio, and body mass index) and it’s starting to bug me now. That the researchers have found that the placebo effect, in the form of advice, can affect such things as body fat is surprising but what’s really bugging me is how it can. I started by trying to think of fairly obvious things.

Like, did being told how much exercise they were actually doing have an effect on their outlook when it came to other aspects of health? Perhaps they were feeling virtuous by dint of their new found awareness of just how active they were and this had a knock-on effect in things like food choice. When I re-read the blog post, I saw that this had been referred to already (“maybe the cleaners changed their behaviour, or their diets, in ways that the researchers didn’t pick up”), so that idea is already out there. I thought that perhaps someone would repeat the study and this time measure food consumption in both groups as well as exercise levels and so be able to at least give us a clue as to whether that idea has legs. It turns out that the researchers had asked the subjects to self-report diet so if their self-assessments were accurate and no dietary changes were made [comment on the Bad Science blog from someone who blogs as ‘A Healthy Distrust’], then that’s idea #1 out of the frame.

Next up, we have the idea that both groups initially underestimated the amount of exercise they were taking, and that the group that was given the presentation not only subsequently had a better understanding of how much exercise they were actually getting but also increased their activity levels. Whereas the control group simply underestimated their level of activity at both the beginning and end of the study and made no conscious or unconscious increase in their levels of activity. Not having access to the full paper made it a little tricky to evaluate this idea. However, it seems like a non-starter. The abstract states that “actual behavior did not change”, so it looks like the authors may have monitored actual levels of activity. I had thought I was being really clever coming up with a hypothesis that their activity levels and awareness of their activity levels improved in tandem as there seemed to be parallels with Unskilled and Unaware of it by Kruger and Dunning (it’s sort-of a paper about improving competence in order to improve awareness of competence/incompetence – pdf). It seemed neat that telling a group of people how their work was exercise, and how much exercise they were getting by working, could increase both their level of exercise and their understanding of how much exercise they were getting – but if the authors monitored the levels of activity in the two groups and these did not change then it looks like it’s a duff idea, however neat it seemed at the time.

Next we have something I find particularly unlikely. Perhaps just thinking about how much exercise you take and about how fit you are has some kind of physical effect in itself. Perhaps thinking about exercise (or telling yourself you are working hard, rather) releases chemical messengers in your brain that have a “knock-on effect” in your body. Maybe (and I’m speculating wildly here), thinking about exercising causes neurotransmitters and hormones to be released that mobilise fat cells in order that your body can utilise the energy and this helps to reduce your BMI, levels of body fat, weight and waist-to-hip ratio. The only thing I’ve missed in that list is blood pressure, which I wouldn’t be surprised to see a drop in if you’ve just told someone they meet the surgeon-general’s recommendations for exercise. I know my first thought would either be “phew, that’s a relief” or “that’s bollocks, I never exercise” – if it were the former then I imagine my more relaxed outlook to that aspect of my health could have a beneficial effect on my blood pressure. The blood pressure readings don’t seem as hard to explain as the other measurements, but the idea that just thinking about exercises causes a release of chemical messengers that melts away your fat seems a bit bloody dubious to me. I mean, what happens to the energy for one thing? We all know energy is not created or destroyed, but is changed so what does it change into if it is released from fat cells but is not used in order to fuel higher activity levels? Heh – maybe we should measure the participants’ temperatures as well, see if they’re heating themselves by burning fat through the power of thought. Or use a surrogate marker such as number of days when jumpers and cardigans are worn.

The killjoy moment: someone emailed Ben Goldacre to point out that “the researchers seem not to have corrected for clustering in their data […] this looks like a pretty valid criticism, and might reduce the statistical significance of the findings, without the raw data nobody can tell by how much.” If it turns out that the findings aren’t statistically significant then there’s no effect to ponder over and I’ve been wasting my time thinking about how revising someone’s impression of their activity level can improve such physical measurements as body fat, weight and BMI. It was fun thinking about it though, and there are certainly worse ways to waste your time. [This is the comment].