Does a firm mattress help back pain? The evidence points to medium-firm

"Get a firm mattress for your back" is one of the most durable pieces of household health advice — and one of the least supported by evidence. As the authors of a landmark trial put it, a firm mattress is commonly believed to be beneficial for low back pain, although the evidence for that recommendation was lacking [s1]. When the question was finally tested properly, the answer landed in the middle: not the firmest surface, but a medium-firm one.

The trial that moved the recommendation

The key study, published in The Lancet in 2003, was a randomised, double-blind, controlled, multicentre trial of 313 adults with chronic non-specific low back pain — people who ached while lying in bed and on getting up, but had no pain radiating down a leg [s1]. Firmness was set on a standardised European scale where 1.0 is firmest and 10.0 softest. Participants were assigned either a firm mattress rated 2.3 or a medium-firm mattress rated 5.6, and assessed at the start and again at 90 days [s1].

The medium-firm group did better on every primary outcome. Compared with the firm mattress, medium firmness roughly doubled the odds of improvement in pain while lying in bed (odds ratio 2.36, 95% CI 1.13 to 4.93) and in disability (OR 2.10, 95% CI 1.24 to 3.56), with a similar but less certain signal for pain on rising (OR 1.93, 95% CI 0.97 to 3.86) [s1]. Across the whole study period, the medium-firm group also reported less pain on rising (P = 0.008), and non-significant trends toward less daytime back pain (P = 0.059) and less pain while lying in bed (P = 0.064) [s1]. The trial's conclusion was blunt: a mattress of medium firmness improves pain and disability in this group [s1].

That the study was double-blind matters. Participants could not see the firmness rating and were not told which mattress they had received, which strips out much of the expectation effect that plagues bedding research — a field otherwise dominated by manufacturer testimonials.

What the wider evidence says

A single trial, however clean, is not a rule. A systematic review published in Sleep Health in 2015 gathered the controlled trials on mattress design published since 2000 — 24 studies in all — and rated their methodological quality as moderate to high on the PEDro scale [s2]. Its verdict aligned with the Lancet trial. A mattress subjectively rated medium-firm, or one that is custom-inflated so the sleeper can self-adjust it, was optimal for promoting sleep comfort, sleep quality, and spinal alignment [s2]. The review found the evidence insufficient to settle side questions such as the ideal mattress temperature [s2].

The convergence is the useful part: an independent trial and a review of two dozen studies point the same way, away from maximum firmness and toward the middle of the range — or toward an adjustable surface the sleeper can tune.

The caveats worth keeping

Neither source turns "medium-firm" into a precise prescription. Firmness ratings are not standardised across brands; the Lancet trial's 5.6 on a laboratory scale does not map cleanly onto a showroom label reading "medium" [s1]. The systematic review leaned partly on subjective firmness categories, which vary from one sleeper to the next [s2]. Body weight, preferred sleeping position, and personal comfort all shift where "supportive" sits for a given person — a point the "self-adjusted" finding underlines [s2].

The trial also studied a specific group: adults with chronic, non-specific back pain and no leg symptoms [s1]. It does not speak to sciatica, to a fresh acute injury, or to back pain with red-flag features — leg weakness, groin numbness, or bladder or bowel changes — which need medical assessment regardless of what a person sleeps on.

The bottom line

The evidence does not support buying the hardest mattress on the floor for a bad back. Two independent lines of research favour medium firmness, or an adjustable surface, over both extremes [s1][s2]. For a shopper, that reframes the task: rather than chasing firmness as a proxy for support, the better test is a mattress that keeps the spine comfortable in your own sleeping position — and, given how much individual comfort varies, one bought with a trial period and a return policy.