A back support helped in one trial. Then Cochrane pooled eight of them and found a gap

Two weeks ago I wrote here about a French trial in which a soft lumbar belt eased back pain more than usual care did. On September 2 the Cochrane Collaboration published its review of that entire field, and the entire field turned out to be eight studies and 501 people.

What the review went looking for

The review asked a much wider question than any single belt. A team led by Chiara Arienti set out to assess every kind of assistive technology used for chronic low back pain: soft and rigid braces, belts, supports, and the walkers, canes and gait aids that people lean on when their back stops them getting around. They searched five databases plus the trial registries up to January 2025 (Cochrane Database of Systematic Reviews, 2026).

Eight randomised trials came back. Five hundred and one participants between them, all adults whose back pain had lasted at least a year. And every single one of those eight studies looked at lumbar supports. Not one eligible trial had ever tested any of the other devices, the mobility aids and gait aids that get handed out constantly in ordinary clinical practice. That half of the question has simply never been asked in a study that met the bar.

Where results could be pooled, they came out small and unsteady. A support worn on its own, compared with nothing, may make little to no difference to pain at three months, and little to no difference to what a person can actually do, on the strength of a single study of 107 people. Worn on top of anti inflammatory medication, supports did shift short term pain by an amount the reviewers describe as a small reduction, across two studies and 149 people. Add a support to education and exercise, or to routine physiotherapy, and the findings scatter so widely across trials of 25 and 41 participants that the review says plainly it does not know. Most of the studies carried a high or unclear risk of bias. The certainty of the evidence was rated low or very low the whole way through.

The column that stayed empty

Here is the part that stopped me. The reviewers had listed, in advance, the outcomes they most wanted: pain, disability, quality of life, whether the person themselves felt the treatment had worked, falls, adverse events, and how many people quit because of one.

None of the eight studies reported whether participants felt the treatment had succeeded. None reported falls. None reported adverse events at all, or withdrawals due to them. After eight trials and decades of these things being worn every day by people all over the world, the harms column is blank. Not reassuring. Blank.

The arc this sits in

Follow the line and it is a strange one. The American College of Physicians published its guideline on noninvasive treatment of low back pain in 2017, and for chronic pain its first line list runs through exercise, multidisciplinary rehabilitation, acupuncture, mindfulness based stress reduction, tai chi, yoga, motor control exercise, progressive relaxation and several others (Annals of Internal Medicine, 2017). Lumbar supports appear nowhere on it. Cochrane registered a protocol to look properly at assistive devices in 2024. Then in August 2026 a multicentre French trial reported that a belt beat usual care on pain and disability, while quality of life showed no difference between groups and physical activity did not rise in the belt group either (JAMA Network Open, 2026). Two weeks later this review lands and says there is insufficient evidence to support routine use.

What ties those points together is not that supports fail. It is that the question keeps being asked at the level of pain intensity, and almost never at the level of what the device is doing to the person wearing it. The French trial found people hurt less and moved no more. Cochrane found nobody counted the falls. Different studies, same blind spot.

What I notice in the room

A woman came to a session last February wearing a support she had kept on most of every waking hour for eleven months. Her physiotherapist had suggested it during a bad flare and she had never taken it off. When I asked her to notice how her lower back moved as she turned to look behind her, she paused for a long time and said she couldn't tell. Not that it hurt. She could not tell.

I don't know whether the belt caused that, and one person in one room is not evidence. But it is the thing the trials are not measuring. A region you cannot feel is a region you cannot organise well, and in the Feldenkrais Method® that is close to the whole point: Awareness Through Movement® lessons work by making small differences noticeable again in places that have gone quiet, which is why so much of the work happens slowly, in small range, with the effort deliberately taken out. Chronic lower back pain rarely stays only mechanical for long, and the way sitting shapes the lower back is a good example of how a region can gradually drop out of a person's own map of themselves.

None of that is an argument against wearing a support. Feldy doesn't replace clinical care, and whether a brace belongs in your week is a conversation for you and the clinician who knows your back. Gentle awareness work sits alongside what your physiotherapist gave you rather than competing with it.

The question worth carrying is a quieter one, and the trials have not touched it yet. When the support comes off in the evening, what does your back feel like underneath?

Eight trials. Five hundred and one people. The falls column still empty.