Ever since high school, when troubles were weighing me down, I went on a run to sort things out. Even if I didn’t come to grand conclusions during those miles, afterward I had the feeling that despite the bad grade or lost love, I was still alive and free.  Exercise lifted my mood.

So it was interesting to see a review in The Cochrane Library by Gary Cooney and Kerry Dwan assessing whether exercise could treat clinical depression. They examined all the studies across the past decades that tested exercise’s effectiveness for treating depression. The results were mixed and the studies varied in quality, with some having important flaws.

When Cooney and Dwan included all 35 potential studies comparing exercise to no treatment, they found that exercise provided a modest benefit for depressed patients. However, when they included only the 6 studies that made every effort to minimize bias, exercise’s benefits were small and statistically insignificant. That’s far from the slam dunk I was hoping for.

There’s a good rationale for why exercise might alleviate depression. One common symptom of depression is decreased physical activity. Exercising, by definition, ramps up activity, so it resolves a major aspect of depression. Further, there’s evidence that it can boost healthy brain activity. A leading model of depression suggests that brain chemicals that affect mood—serotonin, endorphins and the like—are out of balance. Exercise could help restore balance, for example by increasing brain-derived neurotrophic factor—a chemical responsible for nerve cell growth. 

Exercise certainly has appeal as a treatment. It doesn’t require taking any pills or making appointments. Just lace up your shoes and get moving. But here lies the risk: we assume it works because it makes sense—the naturalistic fallacy. Because we’re so quick to buy into it, we may overlook the possibility that it doesn’t work.

Researchers have been trying to show that exercise can treat depression for years, and many of the studies produced favorable results. The 35 studies comparing exercise to no treatment showed that exercise reduced depression ratings by a little over half a standard deviation, a moderate effect size. Further, when they compared exercise to common treatments, psychotherapy or antidepressants, exercise fared just as well. So far, exercise looks like a great option.

The problem is that most of the studies failed to control for three sources of bias: treatment concealment, intention-to-treat analysis and blinded assessment. Put plainly, in quality studies, researchers should remain unaware of what treatment the patient receives during treatment and analysis. Also, all participants who start the study should be accounted for in the analysis. For example, if half of the exercising participants drop out of the study, there may be a problem with exercise. If you only include the half that finished in your analysis, you miss the true story.

Doctors can confer benefits on patients just by expecting them to improve, a form of the placebo effect, so you want them blind to which treatment each patient receives. As a patient, you want a doctor with a large placebo effect because you’ll likely feel better regardless of the treatment. As a researcher, you want a doctor with no placebo effect. You want to be sure that exercise—not the doctor—helped patients.

Only six studies controlled for each source of bias. Those six still showed improvement of depression following exercise compared to no treatment, but the effect was only a fifth of a standard deviation. That’s tiny. Statistically, there was no difference. Exercise isn’t looking so hot, now.

So while it’s tantalizing to home in on studies that show a modest but consistent benefit, it may be too early to conclude that the positive results are due to exercise. Since many people may assume that exercise is beneficial, they may tilt the results of studies on exercise to be more positive. To determine how well exercise works as a treatment, we need more high-quality research.

Until then, feel free to go for a run if it makes you feel better.

image credit: @mattyrunrun

You are reading

You, Illuminated

Born Late: The Trouble With Due Dates

A classic study suggests what we get wrong about estimating delivery.

How the Brain Processes Keeping up With the Joneses

The brain may use two systems for social comparisons

Long Term Effects of Marijuana on the Brain

Part 3 exploring how marijuana alters brain function: long-term effects