Five Stretches Will Not Fix Your Desk Job

The premise does not hold

Start with the causal claim, because everything else is built on it. If sitting at a desk causes back pain, that should be visible in the workers who sit the most. A 2010 systematic review went looking, screening 2,766 citations down to 24 studies, five of them high quality, including three prospective cohorts. It assessed them against the Bradford-Hill criteria, which is the standard framework for deciding whether an association is causal.

It found strong, consistent evidence of no association between occupational sitting and low back pain, moderate evidence of no dose-response relationship, and no statistically significant result on temporality. The conclusion is unusually direct for a review of this kind: it is unlikely that occupational sitting is independently causative of low back pain in the populations of workers studied. No dose-response is the detail that matters most. If sitting accumulated damage, the people sitting longest would hurt more. They do not, reliably.

This does not mean desk workers are fine, and it does not mean posture is irrelevant to how you feel in a given hour. It means the specific story, that hours of sitting are steadily doing structural harm you then have to undo, is not supported.

Stretching is a real tool aimed at the wrong job

The second plank is that sitting leaves muscles shortened and stretching restores their length. In 2025 a Delphi panel of international stretching researchers worked through the evidence across eight topics and published where they could and could not reach consensus. They agreed stretching improves range of motion and reduces muscle stiffness. They also reached consensus that stretch training does not contribute substantively to muscle growth, does not serve as an all-encompassing injury prevention strategy, does not acutely enhance post-exercise recovery, and does not improve posture.

The last one is the plank. Stretching for posture correction is the single most common reason a desk worker is handed a mobility routine, and the researchers who study stretching declined to endorse it.

The structural picture is similar. A 2023 meta-analysis of 19 studies looked at whether static stretching training actually changes muscle architecture. It found increases in fascicle length at rest that the authors themselves classify as trivial, and only when high stretching volumes and intensities were used. No change in muscle thickness at ordinary doses. Whatever a daily stretch is doing for you, remodelling the tissue is mostly not it. The better-supported explanation for feeling looser is a change in how much stretch you tolerate.

So stretching keeps a legitimate place. If you want more range at a joint, it is the direct tool for that, and range is worth having. What it will not do is correct your posture, prevent injury on its own, or reverse a structural debt that the sitting was not creating.

What actually moved the numbers

Here is the useful part. There are trials in exactly this population, and they point somewhere specific: not at the position you hold, but at how long you hold any position, and at how much load the body is used to.

A 2021 trial followed office workers who already had neck or low back pain and randomised them to active breaks, to prompted postural shifts, or to a control group, tracking them monthly for a year. Median time to recovery was one month in the intervention groups against two in the control. Recurrence over the follow-up was 44 percent in the control group, 21 percent with active breaks, and 18 percent with postural shifts. Adjusted for biopsychosocial factors, both interventions were protective against recurrence.

Read the design carefully, because the scope is narrower than the headline suggests. These were workers who already had pain, so this is evidence about recovery and recurrence, not about preventing a first episode. What the interventions had in common was not a posture. It was the interruption itself.

1. Interrupt the position

   Active breaks and postural shifts

   Recurrence 44% to 21% and 18%, with recovery time halved, in office workers with existing pain.

2. Walk

   An individualised, progressive walking programme

   Significantly reduced recurrence of low back pain over 12 months in a 2024 Lancet trial, and was cost-effective.

3. Build capacity

   Exercise, as prevention

   Risk of an episode 0.65 and of sick leave 0.22 for exercise alone. Leisure-time activity is associated with 11 to 16 percent lower risk of chronic low back pain.

The walking trial is worth sitting with because it is the least glamorous intervention imaginable and it worked. Participants were given an individualised, progressive walking and education programme, and low back pain recurrence dropped significantly over 12 months at a cost per quality-adjusted life year that made it comfortably cost-effective. The honest caveat, which the paper reports plainly, is that the walking group logged more lower-limb adverse events than the control group, 100 against 54. Progressive means progressive.

And the prevention picture is consistent. A 2016 meta-analysis found exercise alone reduced the risk of a low back pain episode to 0.65 and the risk of taking sick leave to 0.22. A separate review put leisure-time physical activity at an 11 to 16 percent reduction in chronic low back pain risk, with the authors cautioning that the underlying studies have limitations and the effect needs confirming.

Two practical footnotes on the break research, because a strategy is only as good as the version you can actually run. A follow-up analysis of the same trial looked at who benefited most and found the effects were moderated by ordinary working conditions: how many hours people worked, how demanding they found the job psychologically, and, for low back pain specifically, whether their chair had lumbar support. This is not a protocol that lands identically on everyone.

And the honest ceiling on what a break programme shifts: a 2025 study of short active breaks in office workers found sedentary time during working hours down by 24.4 minutes at one year, with physical activity up by a comparable amount. The authors state plainly that methodological limitations prevent attributing the effect definitively to the intervention. Twenty-four minutes is a real change and a modest one, and it is a more useful number to plan against than a promise that five movements fix most of it.

What the same evidence says does nothing

The prevention meta-analysis is more valuable for its null results than its positive ones, because the null results are the things people buy.

• Back belts: risk ratio 1.01. No protective effect.
• Shoe insoles: risk ratio 1.01. No protective effect.
• Education on its own: risk ratio 1.03 for episodes. Being told how to sit, in the absence of anything else, did not change what happened.

That third one deserves the emphasis. The ergonomics briefing, the posture diagram, the reminder to keep your screen at eye level: as a standalone intervention, none of it changed outcomes. It was exercise, alone or combined with education, that did.

So the practical version is short, and it is not a stretching routine. Break up long stretches of sitting, and treat the interruption itself as the active ingredient rather than the shape you move into. Walk, progressively, and expect to build up rather than start at volume. Strength-train the whole body on a schedule you will actually keep, because that is the intervention with the strongest prevention evidence behind it. Stretch if you want range at a particular joint, which is a real and reasonable goal, and stop asking it to fix your posture.

None of this is medical advice, and none of it replaces assessment by a clinician who can examine you. Pain that radiates down a leg, that follows a specific injury, or that comes with numbness, weakness, or any change in bladder or bowel control is a reason to be seen promptly rather than to start a routine.