Exercise and depression: what Cochrane actually says

Every few weeks a headline turns up claiming that exercise cures depression. It usually comes from a single study, and it usually overstates things. So when Cochrane — the people whose whole job is pooling every study and squinting at it — updates its review on exercise and depression after thirteen years, it is worth reading slowly. Spoiler: the news is good, just not the news you were sold.

What they actually looked at

The review was published on 8 January 2026 and updates the earlier versions from 2008 and 2013. It pools 73 studies and at least 4,985 adults with depression, of which 37 studies supplied enough data for the combined analyses. This is not one more trial: it is an attempt to answer the question with everything published laid out on the table.

The headline conclusion, in Cochrane's own careful wording: exercise may moderately reduce symptoms of depression compared with doing nothing. Note the "may" and note the "moderately". Cochrane does not hand out adjectives lightly.

The awkward bit: better studies, smaller effect

Here is the part that rarely makes the headlines. When the authors restricted the analysis to the best-conducted studies — those that properly concealed which group each participant went into and assessed outcomes blind — that moderate effect stopped being statistically significant.

In plain terms: some of what we see might be the enthusiasm of knowing you are in the group that does something, rather than the exercise itself. That does not mean it is useless. It means the real benefit is smaller and less certain than we would like, and that anyone promising you a tidy number is selling something. It is the same test we applied when working through which wellbeing trends are worth your time: look at who is measuring, how, and what survives once you strip out the noise.

Better or worse than therapy?

The review also compared exercise with psychological therapy and with antidepressants, and found little to no difference between them. That sounds spectacular, which is exactly why the footnote matters: the comparison rests on a handful of fairly small trials. "We found no difference" and "we have shown it makes no difference" are not the same sentence, however similar they look written down.

What was clear is safety: unwanted effects were uncommon. What was not clear is the long term, because most studies did not follow participants beyond the end of the programme. We know a fair amount about the intervention weeks and very little about six months later.

The good news: no need to wreck yourself

If there is one thing to take from all this caution, it is that nowhere does it say you have to suffer. Light and moderate intensity comes out at least as well as hard, and that changes the entrance fee completely. We are not talking about hill reps at seven in the morning. We are talking about walking, about pedalling without rushing, about moving enough to notice you are breathing.

Which happens to be the only sustainable version anyway. A plan that grinds you down gets abandoned in three weeks, and an abandoned programme has a measured effect of precisely zero. If a small target helps you hold on to something, the Hades Challenge is completed with 50 accumulated kilometres and no deadline, and walking counts exactly as much as running. One kilometre at a time, at your pace.

Exercise may be moderately effective compared to no therapy for reducing symptoms of depression.

Cochrane review "Exercise for depression", January 2026

And now, the important part

This is not a medical article and a kilometre challenge is not a treatment. Depression is an illness, it is diagnosed and it is treated, and the person who should tell you what you need is a healthcare professional, not a sports blog or an app with medals. If you are going through it, talk to your doctor or a psychologist: exercise can sit alongside treatment, but it does not replace it.

That said, few things in health carry this much accumulated evidence, this few side effects and this little cost as going out for a walk. Even if the effect is moderate. Even if the scientists say so under their breath.