People with depression drop out of exercise less often than control groups

Exercise has established efficacy in improving depressive symptoms, but a treatment only helps if people can stick with it. Doubts about adherence are a common reason to hesitate before recommending, or attempting, an exercise programme during depression. Dropout from randomised trials also matters scientifically, because people leaving studies threatens the validity of the evidence base itself. This review set out to measure how many adults with depression actually drop out of exercise trials, and to identify what predicts who leaves and who stays.

Three authors identified randomised controlled trials from a recent Cochrane review and ran updated searches of major electronic databases covering January 2013 to August 2015. They included trials of exercise interventions in people with depression, spanning both major depressive disorder and depressive symptoms, that reported dropout rates. In total, 40 trials qualified, reporting dropout across 52 exercise interventions in 1720 people with depression, with an average age of 49.1 years and 72% women. The team then pooled dropout rates and ran further analyses to find what predicted higher or lower dropout.

The adjusted prevalence of dropout across all studies was 18.1%, and 17.2% in trials restricted to major depressive disorder. Among people with major depressive disorder, higher baseline depressive symptoms predicted greater dropout. Interventions supervised by physiotherapists or by exercise physiologists both predicted lower dropout. In a comparative analysis of 29 trials, dropout was actually lower in the exercise arms than in the control conditions, and the authors conclude that exercise is well tolerated by people with depression.

This strengthens the case that exercise is not only effective against depressive symptoms but also feasible as a treatment, answering a practical objection that adherence would collapse in this group. The authors argue exercise works best as a treatment when delivered by healthcare professionals with specific training in exercise prescription, since supervision was the clearest protective factor against dropout. What the review cannot say is whether adherence holds outside the structure of a trial, over longer periods, or among people who would never volunteer for a study in the first place.

For anyone managing low mood who worries that starting to exercise would be setting themselves up to fail, this is evidence that most people in trials kept going, and that those assigned to exercise quit less often than those who were not. It is also a reason to think structure and supervision matter: programmes run by trained professionals held onto more of their participants. People with more severe symptoms found it harder to stay in, which is worth knowing rather than a judgement. None of this speaks to which type of exercise works best.