Anecdotal benefits have been reported by wearers but studies comparing these bracelets with 'dummy' versions have produced mixed results.

The current British Medical Journal study found a significant reduction in pain scores among 65 wearers.

The Peninsular Medical School team said the effect could be real or down to the individual's faith in the treatment.

'Real' benefit

The authors also emphasised that the benefits were in addition to existing treatments, which should not be suddenly stopped without discussion with their doctor.

Also, high strength magnets (170mTesla or more) seemed to be needed to have any effect on pain.

GP Dr Tim Harlow and colleagues recruited 194 patients aged 45-80 years with osteoarthritis of the hip or knee from five rural general practices in Devon.

The patients were given one of three bracelets to wear for 12 weeks - a standard strength magnetic bracelet, a weak magnetic bracelet, or a non-magnetic 'placebo' bracelet.

The patients were asked to rate their pain using a recognised scoring scale.

All three groups reported less pain when wearing the bracelets.

But the largest reductions in pain scores were reported by the patients wearing the standard strength bracelets.

The results for the weak magnet group were similar to those of the dummy magnets, suggesting that the magnetic strength of the bracelet is important.

Placebo effect

Dr Harlow and his team, who were funded by the Arthritis Research Campaign, said more research was needed to confirm their findings.

They did checkthat factors such as use of painkillers and patients' beliefs about the type of bracelet they were testing had not affected the results.

They said: "We cannot be certain whether our data show a specific effect of magnets, a placebo effect, or both.

"Whatever the mechanism, the benefit from magnetic bracelets seems clinically useful."

A spokeswoman for the Arthritis Research Campaign said: "We funded this study because we wanted to establish if there was any evidence for the claims made on behalf of magnetic bracelets; and we didn't want the public to waste their money on devices that didn't work.

"Results appear to show that wearing a magnetic bracelet does reduce pain in people with hip and knee osteoarthritis although it is still unclear whether this effect is due in some part to the placebo effect.

"As magnetic bracelets are quite cheap, between £30 and £50, and safe, people with osteoarthritis might want to consider wearing them as part of their self-help regime."

However, a spokesman from Arthritis Care said: "Clear evidence of the efficacy of magnetic bracelets as a means of treating the symptoms of arthritis is yet to be established.

"This is due mainly to the lack of large-scale clinical trials undertaken in this regard.

"As a consequence, Arthritis Care does not recommend the use of magnetic bracelets for this purpose, though we would welcome a more robust and expansive trial of this treatment as a means of providing firm evidential grounds for optimism."

About 760,000 people in the UK have osteoarthritis.