Walking May Match Therapy for Depression, and Beat Antidepressants. A 218-Trial BMJ Analysis Says So

The Story Everyone Knows

Depression is treated with two tools: pills and talk, and when a doctor mentions exercise, it arrives as an afterthought, the gentle suggestion tacked onto the prescription, the one patients nod along to and then forget by the time they reach the parking lot. Move a little, get some sun, it can't hurt. Nobody tells a depressed patient that the morning walk might be the treatment itself.

That hierarchy felt obvious because the evidence was lopsided. Antidepressants had hundreds of randomized trials behind them; psychotherapy had decades of manuals and meta-analyses. Exercise had scattered, underpowered studies, many measuring mood in healthy volunteers rather than in people with diagnosed depression, and the guidelines reflected that asymmetry: exercise was the garnish, never the main course.

The assumption was never that exercise does nothing, which everyone concedes; the assumption was that what it does is small, unreliable, and certainly not comparable to an SSRI or twelve weeks of cognitive behavioral therapy, and that is the assumption a team publishing in the BMJ in February 2024 set out to dismantle with the largest evidence base anyone had assembled.

The Study

Noetel's team pooled 218 randomized trials and 14,170 participants into a single statistical model. Eligibility was strict: every trial enrolled people meeting clinical cutoffs for major depression, and the exercise had to be a real program, not a one-off stroll. Because few trials tested exercise directly against therapy or drugs, the researchers used network meta-analysis, which links treatments through their shared comparison groups to estimate how each would fare against the others, a design that buys breadth at the cost of comparing trials that were never built to be compared.

Compared with active controls such as usual care or a placebo tablet, walking or jogging reduced depressive symptoms with an effect size of g −0.63, drawn from 51 trials and 1,210 participants. Yoga came in at g −0.55 and strength training at g −0.49, while inside the same network, cognitive behavioral therapy alone scored g −0.55 and SSRIs alone g −0.26.

Read those numbers twice. Walking performed about as well as therapy and roughly two and a half times as well as antidepressants alone, with strength training and yoga also outperforming the drugs. Benefits scaled with intensity, with vigorous exercise beating light activity along a clear dose-response curve, and strength training and yoga proved the most acceptable treatments, with participants about half as likely to drop out as controls.

The Arithmetic Nobody Did

Walking's effect of −0.63 clears the paper's minimum clinically important difference of −0.20 by more than three times, while the SSRI estimate of −0.26 clears it by a hair. By this analysis, a walking program is not a marginal gain but a large one.

The largest effect in the whole network belonged to dance, at g −0.96, the number that got the headlines. But it rested on five studies and 107 people, most of them young women, while the walking estimate rested on 51 trials and 1,210 participants: more than eleven times the evidence base. The most spectacular result in the analysis is also the least trustworthy one. The most credible result is a brisk walk.

The Strongest Counterargument

Now the part the headline leaves out. Of the 218 trials, exactly one met the Cochrane criteria for low risk of bias. One. Blinding participants to an exercise program is essentially impossible, so expectancy effects could be inflating every estimate in the network. The authors graded the evidence as low confidence for walking and very low for everything else. In plain language: the effects look real, but the trials behind them are mostly weak.

The SSRI comparison deserves its own caution. The review was designed to find exercise trials, not antidepressant trials, and the authors explicitly warn that their drug estimate should not replace reviews built to evaluate pharmacotherapy. The exercise-to-SSRI comparison is indirect, flowing through shared control arms, and some treatment nodes drew on systematically different patients: severe depression was more common in the SSRI studies, while older adults clustered in tai chi, and the paper's own incoherence checks found direct and indirect evidence disagreeing on at least one comparison. The network is an elegant statistical machine, but it still compares trials that were never designed to be compared.

Then there is the practical problem the trials barely address, which is adherence: telling someone with anhedonia and leaden paralysis to start jogging is easy to prescribe and brutally hard to follow, and dropout from a study is an imperfect proxy for adherence in the wild, where no researcher is checking in. Exercise remains the treatment that people with depression are least likely to start precisely when they need it most, which is the cruel paradox at the heart of every prescription: the intervention works best for those who can still get themselves out the door.

What We Didn't Prove

This analysis does not prove that exercise replaces antidepressants or therapy. It does not prove the benefits last: measurements ran out to about six months, and maintenance beyond that is uncharted. It does not identify the mechanism, though the authors gesture at social contact, mindfulness, self-efficacy, green space, and neurobiology. And it cannot rule out expectation: a person who knows they are in the exercise arm expects to feel better, and people who expect to feel better sometimes do.

It also does not prove that one exercise fits everyone: strength training worked better for women, yoga and qigong for men, yoga for older adults, and strength training for younger people, which is why the review argues for prescribing exercise the way drugs are prescribed: matched to the patient, dosed by intensity, and followed up.

The Bottom Line

Depression treatment carries a quiet embarrassment: only half of people with depression in high-income countries receive any treatment, while the gold-standard options are expensive, scarce, and resisted by the very patients they are meant to help. Against that backdrop, a free treatment with no pharmacy, tested in 51 randomized trials of walking alone, matching therapy's effect and doubling the drug's, is not a curiosity. It is an intervention whose professional status has lagged its evidence.

The evidence is weaker than the headline and stronger than the garnish. Both can be true.

What You Can Do

Talk to your clinician about exercise as treatment, not decoration. The trials tested structured programs rather than vague intentions: scheduled sessions over several weeks, with a defined modality and intensity, designed with your clinician so progress can be measured and adjusted like any other prescription.

Start with walking. It had the largest evidence base and one of the largest effects of anything in the network. Intensity matters, so work up toward a brisk pace rather than an amble, though even light walking cleared the clinically meaningful threshold.

Consider strength training or yoga if sticking with it worries you. These two modalities had the lowest dropout rates, roughly halving the odds of quitting compared with control conditions.

Do not stop prescribed medication on the basis of this study, because the comparison with SSRIs was indirect, the confidence ratings were low, and the authors position exercise as a core treatment alongside existing options, not a substitute for them.