Nipun Shrestha and colleagues gathered 34 trials with 3,397 participants for the Cochrane review, all of them from high-income countries. Sit-stand desks reduced sitting at work by 100 minutes a workday at follow-ups of up to three months and by 57 minutes at three to twelve months. Counting sitting outside work as well, the total came down by 82 minutes a day.
Shrestha rates the evidence for sit-stand desks as low quality, finds no evidence at all about what happens beyond a year, and calls the effects of every other approach mostly inconsistent: workplace policy changes, information and counselling, multi-component programmes. The review found no study at all on standing meetings or walking meetings.
Zhou ran a network meta-analysis over 23 trials and 1,428 office workers, which lets the options be ranked against each other. A sit-stand workstation on its own was −1.10 as a standardised mean difference against a typical desk. Adding reminders about breaks and posture took it to −1.49, and a multicomponent intervention to −1.50. Certainty of evidence, again, low.
Parés-Salomón pooled 11 trials of workplace interventions with digital components and found 29.9 minutes less sedentary time per eight-hour workday. Leppe-Zamora looked only at computer prompt software across 18 trials and 1,164 workers, and found 12.46 minutes less sitting a day measured by accelerometer. Both figures are a fraction of the 82 minutes the desk itself buys.
Bethany Barone Gibbs randomised 27 desk workers who already had chronic low back pain to counselling, a sit-stand attachment, an activity-prompting wristband and cognitive behavioural therapy, or to nothing. Work sitting time fell by 1.5 hours a day. Disability from back pain fell by 50 per cent from baseline in the intervention group against 14 per cent in the control, at p = 0.042. Pain itself, on a visual analogue scale, did not separate significantly.
Silva randomised 38 office workers to a sit-stand desk with a session and prompts, or to their usual conditions, and counted the sitting outside work as well as the sitting at it. After six months they sat 32.97 minutes less and stood 27.88 minutes more at work. Leisure sitting rose by about 23.07 minutes, which did not reach significance, so the compensation is a suspicion rather than a finding.
The two trials here have 27 and 38 people. The four reviews pool other people’s trials, and three of the four rate their own certainty as low.