A lumbar belt eased back pain in a new trial. How much people moved did not change

Most back pain guidelines have said the same thing about lumbar belts for years. The evidence is not strong enough to recommend them. That is the position the authors of a new trial name in their own opening line, and then they report a result that runs the other way.

What the trial did

The study ran across 17 medical centers in France and was published on August 19 in JAMA Network Open. It enrolled 168 adults whose nonspecific low back pain had lasted between one and six months. Mean age was 49, and 60.1% were women. Half were asked to wear a soft, nonrigid lumbar belt for 4 to 8 hours a day during daytime activities on top of their usual care. The other half received usual care alone (JAMA Network Open, 2026).

At 12 weeks, disability on the Oswestry Disability Index had improved by 10.0 points in the belt group and 5.3 points in the usual care group. That puts the belt group 4.7 points ahead, with a 95% confidence interval running from 0.9 to 8.4 points of extra improvement (P = .01). Pain at rest and pain during activity both improved more with the belt. Medication use was lower too: 50.6% of the belt group took pain medication compared with 67.6% of the usual care group, and the gap was widest for anti inflammatory medication, at 18.5% versus 36.5%. No serious adverse events related to the device were reported.

What those numbers carry, and what they do not

The 4.7 point difference sits below the 10 point threshold an international consensus panel proposed as a minimal important change on the Oswestry index for an individual patient (Spine, 2008). The authors answer that with the individual level figure instead: 60.5% of the belt group reached at least a 30% improvement, the other threshold that same panel proposed. Both readings are legitimate. Neither turns this into a large effect.

The trial was also open label. There was no sham belt, and there could not realistically be one. Participants knew whether they had been handed a device. The authors say plainly that placebo, expectation, and nocebo effects cannot be excluded, and that the finding should be read as the overall effect of prescribing a belt in ordinary clinical care rather than the effect of the fabric on its own. The study was funded by Thuasne, which makes the belt, and three authors received personal fees from the company. None of that invalidates a multicenter randomized trial. All of it belongs in how you read one.

Then there are two secondary results that got much less attention than the headline. Overall quality of life on the EQ-5D showed no significant difference between the groups. And physical activity did not increase more in the belt group than in the usual care group.

The thing that did not change

That second one is worth sitting with. If a belt helps mainly by giving someone the confidence to move again, you would expect the belt group to move more. They did not. They hurt less, they took fewer painkillers, and they went on doing roughly what they had been doing before.

A trial cannot tell you why, and the authors do not claim to know. But it matches something I hear often in practice. When a client tells me a support helped, the sentence that follows is almost never "my back got stronger." It is closer to "I stopped guarding it all day." A belt adds nothing to the back itself. What it may do, and I offer this as an observation rather than a mechanism anyone has demonstrated, is give the trunk a boundary from outside so the person stops supplying one from inside. Holding yourself braced all day is expensive, and most of it happens below the level of noticing.

Which points at the real question underneath this trial. Not whether a belt works, but what it is standing in for, and whether that job comes back to you.

Three questions worth raising at your next appointment

Does this trial actually describe my back? Pain that has lasted one to six months is a different problem from pain you have carried for eight years. Mean age here was 49 and baseline disability was moderate, so the results do not stretch to cover long standing or severe pain. It is fair to ask whether the evidence behind a suggestion matches your situation or somebody else's.

How many hours, and what is the plan for stopping? The protocol was 4 to 8 hours a day during activity, not overnight, and the data stops at 12 weeks. There is nothing here about six months or a year of daily wear. A support with a start date and no end date is a different intervention from the one that was tested, and worth naming as such before you begin.

What are we doing alongside it? This is the one I would ask hardest. In this trial the belt lowered pain without changing how much people moved. If a belt buys you a window where things hurt less, the useful question is what goes into that window. Your physiotherapist can tell you what is reasonable to start with, and a belt sits alongside that work rather than standing in for it.

What is worth doing in the body either way

Whether you wear a belt is a conversation between you and your clinician, and there is now a reasonable trial on the table for it. The part that stays yours is what happens inside the support.

The Feldenkrais Method® works on a different question from the one a belt answers. Rather than adding support from the outside, Awareness Through Movement® lessons use small, slow, unforced movements that let the nervous system notice how you are organizing yourself, including the holding you never decided to do. None of it asks you to be out of pain first. You only go where it feels easy, pleasant, and comfortable, and quite often it is doing less that makes the difference.

You have not lost movement. What tends to go first is the trust that moving will help, and support borrowed from the outside can be a genuinely useful bridge while that trust rebuilds. The thing worth watching is whether it stays a bridge.