No sign exercise works better for some anxious people than others

Interest in individualised responses to exercise has been growing within mental health care and psychiatry, with the appealing idea that the right person matched to the right exercise would get far more relief than a generic plan. But personalisation only makes sense if true differences between individuals actually exist, rather than apparent differences that are really just measurement noise and chance. For adults with anxiety and stress-related disorders, nobody had formally tested whether such true interindividual differences in the calming effect of exercise are real. This analysis set out to answer exactly that question before the field invests heavily in precision prescriptions.

The researchers extracted data from a previous meta-analysis of randomised controlled trials and updated the searches across CINAHL, Embase and Medline. That produced 8 arms from 7 original studies, totalling 322 participants with anxiety and stress-related disorders. To test for true interindividual differences, they compared the spread of anxiety-change scores between exercise and control groups, treating change outcome standard deviations as point estimates. Statistical pooling used inverse variance heterogeneity and restricted maximum likelihood models, allowing the team to separate genuine between-person variation in response from the random variation that any trial produces.

Both aerobic exercise and resistance training showed significant anxiety-reducing effects across the pooled trials, confirming that exercise itself works for this group. The headline result, though, was a null: no significant pooled interindividual differences were found for aerobic exercise, and none for resistance training either. In plain terms, the data gave no convincing sign that some people are true responders while others are true non-responders. The spread of outcomes in the exercise groups was not meaningfully wider than in the control groups, which is what genuine individual response differences would be expected to produce.

This finding pushes back on the precision-exercise trend in mental health care, at least for anxiety and stress-related disorders. The authors conclude that clinical practice can focus on general population physical activity guidelines rather than aiming for highly specific, individualised recommendations, and that research energy might be better spent on how to motivate patients to meet those guidelines at all. The evidence base remains small at 322 people across 7 studies, so subtle individual differences cannot be ruled out, and the question could be revisited as more trials accumulate.

If you deal with anxiety and have wondered whether you need the right kind of exercise, tuned to your physiology, before it will help, this analysis suggests otherwise: both aerobic work and lifting reduced anxiety, and there was no evidence that the benefit varies meaningfully from person to person. That is quietly liberating, because it means the barrier is not finding a perfect personal formula but simply doing enough of whatever activity you will actually stick with. The honest limit is the small evidence base, so this is a reason not to wait for personalisation, not proof that everyone responds identically.