This is the page this site was built to write. If you are searching for a standing desk because your lower back hurts by mid-afternoon, the advice you will find elsewhere is mostly a list of desks with "ergonomic" in the description. That is not an answer. Here is what the research supports, what it does not, and which specifications follow from that.
What the evidence actually says
UT Southwestern Medical Center's summary is the most useful short version we found: "Using a standing desk properly might help relieve back pain caused by excessive sitting," and it cites a 2018 study in which patients using sit-stand desks experienced a 50 percent decrease in low back pain compared with controls. The same article is direct about the limits — a standing desk cannot correct "serious problems, such as scoliosis or a bulging disc" — and warns that standing too much brings its own risks, including leg pain, swelling, and for some patients deep vein thrombosis or varicose veins.
Set against that, the 2018 Cochrane systematic review found only low-quality evidence that sit-stand desks reduce sitting time at all, and no evidence about longer-term effects. Both things are true at once, and the honest synthesis is narrower than either headline: a sit-stand desk is a tool for changing position often. Whether that helps your back depends on whether you actually use it that way.
UT Southwestern's practical guidance is the part worth acting on: transition gradually, use an anti-fatigue mat, keep your ergonomics right, and "change positions, shift your weight, and take short walk breaks every 30 minutes."
The specs that matter for a bad back
- The bottom of the range, not the top. You will spend real time seated. A desk that bottoms out at 25" forces a 5'5" user into raised-shoulder typing, which loads the neck and upper back exactly when the lower back is already unhappy. The UPLIFT V3 (22.6") and FlexiSpot E7 (22.8") are the two here that genuinely go low.
- Programmable presets. The whole mechanism is switching often. A desk with presets gets switched several times a day; a desk with up and down buttons gets switched twice.
- Travel speed. Slow desks do not get used. This is why we excluded crank desks from every roundup on this site.
- Capacity headroom. Not for safety — for smoothness. A frame carrying close to its rating moves less evenly, and a judder on the way up is exactly when people stop bothering.
Why presets matter most
The rule we would give a friend
Set two presets on day one: your measured standing height and your measured seated height. Then move between them on a timer, roughly every thirty minutes, and do not try to stand all day. The desk is only worth what you paid for it if the button gets pressed.
Vari ships four programmable settings, which is one more than the usual three and enough for standing, seated, perched and a second seated position. Preset counts are the least glamorous spec in this category and the one most correlated with a desk still being used a year later.
The desk is a third of the fix
Buying only the desk is the most common way to be disappointed by this category.
- A mat. Standing on a hard floor is its own source of low back pain, and UT Southwestern names an anti-fatigue mat directly.
- Something to sit on. Nobody stands eight hours.
- Getting the height right. A correctly specified desk set to the wrong height is a worse chair.
If the pain is sciatic or disc-related rather than general, those are different problems with different answers — see standing desks and sciatica and standing with a herniated disc.
How we picked
Desks we put on a test rig and measured: none
Plenty of sites imply a lab. We don't have one, and we say so. One desk here is genuinely owner-tested — the UPLIFT V2 that Mike bought and works at daily. Everything else is chosen the way an informed buyer would choose it: read the maker's spec sheet, manual and warranty terms; record the figures that decide whether a desk fits a body and stays still at height; check whether replacement parts are actually sold; then score each pick against a published rubric.
Scores are editorial judgments from that documented research, not lab results, and every product is labeled owner-tested or researched so you can weigh it accordingly. Nobody pays for placement. The rubric and the weights are on the How We Pick page.