When I first saw the coverage of the article appear on Jezebel saying that exercise doesn't help depression, I didn't believe it. I read the press release, and really didn't believe it. And then, I read the article.

Do I believe the article? Yes, I believe that the data are as they say they are. But do I believe that exercise doesn't help depression? Nope. Not a chance. Because that's not what this study says. And in a truly massive failure" of press release and media coverage (some of which was elegantly skewered by Martin Robbins and Tom Chivers), everyone is going to get the wrong idea.


Chalder et al. "Facilitated physical activity as a treatment for depressed adults: randomised controlled trial" British Medical Journal, 2012.

Contrary to some of the statements in the introduction of the paper, there are several meta-analyses which support the effects of exercise in treating symptoms of depression. However, they authors are right, many of the studies have small numbers of people and have extensive exercise interventions. The authors of this study were interested in a milder intervention in a larger group of people: could moderate increases in physical activity buttress depression treatment?

To look at this, they recruited around 360 people who were experiencing a new episode of depression. They assigned half of them to an exercise intervention, and half to control. ALL of them got "normal" treatment, meaning some additionally got talk therapy, some additionally got antidepressants of various types, etc. In the exercise intervention, the group received three meetings with a trained facilitator and 10 phone calls during the year, encouraging them to exercise for 150 minutes per week of moderate to vigorous physical activity in bouts of at least 10 minutes.

The findings showed no effect on depression score as a function of whether the people were assigned to the exercise arm or not. Most people improved in their measures of depression, but there was no difference between the groups.

So...at first blush, it looks like exercise doesn't help depression. But it's really not that simple.

Here's the most important issue: the people in the physical activity intervention did not end up exercising more than controls. The effect was p=0.08 (statistical significance is p<0.05), and the effect was only significant when all three time interventions were considered together. For example, at 4 months from the start of the study, 52% of the physical activity groups were achieving 150 min of moderate to vigorous exercise per day (that's walking and up), compared to 43% of the control arm. The data was self-reported, and the authors also noted that the self-reports didn't line up well with the accelerometry measures designed to measure the physical activity, so it's very possible that even those who reported exercising were not working out as much as they thought they were (not only that, if you're talking to a physical activity facilitator and you know full well the study wants you to exercise, you might talk up your moderate exercise more than you might otherwise).

With adherence like that, no wonder they didn't get an effect. They described good adherence to the study parameters. This means that the people in the exercise arm attended the meetings and phone calls that they were supposed to, but it sure as heck doesn't mean they exercised more. Not only that, I don't think this was the right intervention to really look at. The exercise stuff was designed to fit in your normal day. Walking to your car counts, taking the stairs counts, etc. That's...not an exercise program that can really give you a sense of achievement. In essence, this study asked for the bare minimum of healthy physical activity. Not only that, the study did not discount those who did not exercise, up to 58% of the exercise arm.

What would they have gotten if they had just compared those who exercised against those who didn't regardless of the arm of their experiment? What if they'd looked at depression score as a function of exercise amount? Surely not all of these people were moderate walkers, some must have been more or less athletic than others, did they show differences?

There's another issue as well: most of the patients, but not all, were being treated with antidepressants. It's yet another variable and one they could have, and should have, separated out for analysis. They looked for decreases in antidepressant use in the control and exercise arms (and found no difference), but they did not look for differences in adherence or depression scores in those treated with antidepressants vs not.

In sum, this study did not assess whether exercise helps symptoms of depression. What it assessed was whether someone encouraging you to exercise helped your feelings of depression, regardless of whether you exercised or not.

Is it a well-designed study? Yes, for what it is. Was it well-performed? Not really. Is it well-analyzed? I don't really think so. And was it well reported, in the press release and on? Definitely not. And I'm not the only one who thinks so, the responses of various professionals on the BMJ site itself have pointed out similar flaws.

The fact is, this study does not show that exercise has no effect on depression. It shows that this specific intervention, having someone encourage you to exercise, does not help depression. Not only that, it only shows that having someone encourage you to exercise doesn't help depression scores on top of normal antidepressant treatment. But that sounds a lot less surprising and interesting, now doesn't it.

What we need here is a better study. We need a trial where some receive therapy, some receive antidepressants, some receive exercise (and I'm talking someone being with you when you work out and meeting specific exercise goals that are readily measured and defined), and some receive combinations of these interventions. Then we'll have something that I believe. This study? This doesn't fit the bill.

Chalder, M., Wiles, N., Campbell, J., Hollinghurst, S., Haase, A., Taylor, A., Fox, K., Costelloe, C., Searle, A., Baxter, H., Winder, R., Wright, C., Turner, K., Calnan, M., Lawlor, D., Peters, T., Sharp, D., Montgomery, A., & Lewis, G. (2012). Facilitated physical activity as a treatment for depressed adults: randomised controlled trial BMJ, 344 (jun06 1) DOI: 10.1136/bmj.e2758