Weighted Blanket for Sleep in 2026: 426 Patients, a p-Value of 0.06 and the Honest Case for Buying One Anyway
The best pooled analysis of weighted blankets for sleep landed on a p-value of 0.06. That’s one hundredth of a point short of the line researchers use to say “this worked”, and it’s the number the blanket industry would rather you didn’t read.
In This Article
- Weighted blanket for sleep: what the 2024 meta-analysis actually found
- Half the trials couldn't hide which group was which
- The mechanism is a guess, and a decent one
- The 10% rule nobody has tested
- Who should bother, and who shouldn't
- What to do first, which costs nothing
- A better trial would look different
- The practical stuff retailers skim over
- Why I'd still call it a reasonable gamble
- The verdict, with the maths left in
It comes from a 2024 meta-analysis in Complementary Therapies in Medicine (Zhao and colleagues), which pooled eight randomised trials and 426 patients. Three of those trials measured insomnia. Together they showed a drop of 1.92 points on the scales used, with a confidence interval running from -3.92 to +0.09. In plain English: the blankets probably helped a bit, but the data can’t rule out that they did nothing at all.
And Sunday is when the clocks go back, which is exactly when the sales emails for a weighted blanket for sleep start arriving. So it seems a fair moment to ask what you’d actually be paying for.
Weighted blanket for sleep: what the 2024 meta-analysis actually found
Start with who was studied. Every trial in that review enrolled people with a diagnosed mental health condition, not the general population lying awake at 2am because the neighbours’ telly is on. If you’ve got no psychiatric diagnosis, none of this evidence was gathered on you. That’s not a dismissal. It’s just the first thing to know.
The results, as the authors report them: anxiety fell by a small amount across five studies (a standardised difference of 0.40). Insomnia, across three studies, moved in the right direction without clearing the bar. When the authors dropped one study from the insomnia analysis to see how sensitive the result was, the remaining two gave a clear improvement of 2.78 points (p = 0.001). That sounds like a rescue. It isn’t. A leave-one-out check is a legitimate test, but what it shows here is that the finding hangs on which single study you keep, and with only three studies to begin with, that’s a wobbly place to stand.
The authors’ own conclusion is more careful than most of the coverage it generated. Weighted blankets “may be a safe and effective add-on treatment”, they wrote, but the results “need to be further confirmed” because there were few studies, small samples and large differences between them. You can read the abstract on PubMed.
It helps to translate that anxiety number. A standardised difference of 0.40 is conventionally labelled small, the sort of shift you’d notice on a questionnaire but might struggle to feel on a Tuesday night. Put another way, the blanket nudges the dial rather than turning it. If you’re looking at anxiety that stops you functioning, a nudge is welcome but nowhere near enough, and it’s the reason the authors keep calling it an add-on.
There’s also a gap in what I could check. I wasn’t able to read the full text of a second 2024 review of psychiatric patients (Wong and colleagues, Journal of Psychiatric Research), only its summary line, which describes the synthesised evidence as limited. I haven’t leaned on it for any figures. If you want the detail, go to the paper rather than taking my word for it.
No serious adverse events were reported. That matters, and it’s the strongest card the blanket holds.
Half the trials couldn’t hide which group was which
Here’s the bit that gets lost. According to the same abstract, half of the included studies were rated at high risk of bias, mainly because participants couldn’t be blinded for the outcomes they reported themselves.
Think about what that means. You can’t give someone a fake weighted blanket that feels like a real one. A 7kg blanket is obvious from the moment it lands on your legs. So everyone in the blanket group knew they were in the blanket group, and sleep quality, anxiety and insomnia severity are all things people rate on questionnaires. If you expect to feel calmer under a heavy cover, you’ll probably score yourself as calmer.
That doesn’t make the effect fake. Expectation is a real part of how people sleep, and a bedtime ritual that makes you feel safe is not something to sneer at. But it does mean the true size of the benefit is likely smaller than the headline figures suggest, and nobody can currently say by how much.
Worth saying too: a 2020 occupational therapy review in the American Journal of Occupational Therapy (Eron and colleagues) described the evidence on anxiety and insomnia as sparse. Four years and a handful of new trials later, the verdict has shifted from “sparse” to “promising but thin”. That’s progress, though not the sort you’d put on a box.
The mechanism is a guess, and a decent one
The usual explanation is deep pressure stimulation. Firm, even pressure across the body is thought to calm the nervous system, a bit like a long hug or the tight swaddle that settles a newborn. Manufacturers lean on this heavily, and some talk about raised serotonin and melatonin as if it were settled.
It isn’t. I couldn’t find a trial that measured those chemicals in a way that proves the blanket causes the change, so treat any packaging that does as marketing. The idea is plausible and it fits with how pressure affects the body generally. But plausible is not the same as shown, and a lot of sleep gadgets have gone down the same path.
The 10% rule nobody has tested
Almost every retailer will tell you to pick a blanket weighing about ten per cent of your body weight. So a 70kg adult gets a 7kg cover. It’s repeated everywhere, and I’m guessing here, but it looks like a manufacturer’s convention that stuck rather than something derived from a dose-finding trial. I couldn’t find any UK regulator or trial that sets it.
The studies themselves used a range of weights, so the research can’t tell you whether 7kg beats 5kg for you. Practically, that cuts two ways. If you try one and it feels like being pinned down, it’s too heavy, whatever the chart says. And if you’re buying online, buy from somewhere with a decent returns policy, because the only reliable test is a few nights in your own bed.
Makers also generally say not to use them for babies and toddlers, or for anyone who can’t lift the blanket off themselves. Take that seriously. Check the specific instructions on the one you buy.
Who should bother, and who shouldn’t
Let’s be direct. If you’re sleeping fine and fancy a heavy blanket because it looks cosy, buy it, enjoy it, and stop reading about it. There’s no harm in it and no need for a trial to justify it.
If you have anxiety, restless nights or a diagnosed condition and your clinician is open to it, a weighted blanket is a reasonable thing to try alongside proper treatment. That’s the group the evidence actually covers, and the authors framed it as an add-on, not a replacement. It doesn’t replace cognitive behavioural therapy for insomnia, which is the first-line approach the NHS insomnia guidance points people towards when habits alone aren’t enough.
Where I’d push back is on spending serious money. A heavy blanket is not a £200 problem-solver for chronic insomnia, and the data don’t support pretending it is. If it costs a fortune, you want the return policy more than the brand name. If you’re struggling for weeks, see your GP before you see a shopping basket.
One more point for the sceptics in the room. If a blanket is helping, you’ll usually know within a couple of weeks, not a couple of months. Give it ten to fourteen nights, keep a rough note of how long you take to fall asleep and how many times you wake, and if nothing has changed, send it back. Sunk cost is a poor reason to keep a 7kg object on your bed.
What to do first, which costs nothing
Before the blanket, the boring stuff. The NHS advice on insomnia is built around going to bed and getting up at regular times, keeping the bedroom cool, dark and quiet, and cutting back on caffeine and alcohol in the evening. None of it is exciting. All of it has better evidence behind it than a heavy duvet.
We’ve looked at a few of these already. Our piece on the caffeine cut-off time explains why that afternoon flat white matters more than people think. If dark mornings are what’s wrecking your sleep rhythm, the sunrise alarm clock test covers what actually helps before the clocks go back, and the SAD lamp piece looks at the evidence for winter low mood.
And if you’re tempted by the gadgets that promise to fix things overnight, it’s worth reading what we found about melatonin gummies and the real accuracy of sleep trackers first. Neither is a quick fix.
A better trial would look different
What would settle it? A bigger trial, run on people with ordinary insomnia rather than a psychiatric diagnosis, using a light blanket as the comparison so that nobody can be sure which one is the “real” treatment. It’s not perfect blinding, because heavy and light still feel different, but it’s far better than comparing a weighted blanket with nothing. Add sleep measured by a wrist device or lab rather than a questionnaire alone, and you’d have something worth quoting.
Nobody has to wait for that before trying one at home. But it’s why I’d keep your expectations at “might take the edge off” rather than “will fix it”.
The practical stuff retailers skim over
Heat is the obvious one. A thick, filled blanket traps warmth, and I’m guessing here, but plenty of people will find a heavy cover too hot on a mild night, even in October. It’s less of a problem in a cold flat in January and a real one in a heatwave, which we covered in our guide to sleeping in a heatwave. If you share a bed with someone who runs hot, expect an argument.
Then the faff. Heavier blankets are awkward to wash, and many only go in a machine if the drum is big enough. A removable cover you can launder weekly is worth more than a fancier fill. Weight is spread through pockets of beads or similar filling, so a blanket that bunches in one corner after a few washes has lost most of the point of paying for it.
And if you share a bed, remember a single heavy blanket across two people rarely works. The weight is meant to sit evenly across your body, so two of you pulling in opposite directions defeats it. Two smaller ones, or one each, makes more sense than one enormous one.
Why I’d still call it a reasonable gamble
Which brings me to the contrarian bit. The sleep industry has plenty of products that cost more and prove less. Blue light glasses, for one, failed to show a benefit in the trials we looked at, and we wrote about that in our blue light glasses piece. A weighted blanket at least has trials, a plausible mechanism and no serious harm reported in them. That puts it ahead of most of the pack.
It’s still overrated if you think it’s a cure, and it’s still a decent purchase if you think of it as a comfortable, slightly odd way to feel calmer at bedtime.
The verdict, with the maths left in
A weighted blanket for sleep is a low-risk, plausible and modestly supported add-on for people with anxiety or psychiatric conditions, backed by eight small trials that couldn’t blind anyone. It isn’t a treatment for insomnia, and it certainly isn’t proven for healthy sleepers.
That’s a more useful answer than “it works” or “it’s a gimmick”, and it’s the one the numbers support. A p-value of 0.06 is not a failure. It’s a question mark.
So here’s the one I’d leave you with. If you try one this winter, what will you measure to know whether it’s working: how long you take to drop off, how often you wake, or simply whether you dread bedtime a little less?
