Despite the widespread availability of evidence-based medicine in the western world, ‘alternative medicines’ are still commonly used. Such medicines are usually inspired by pre-scientific medical practices; those which have been passed down through generations. However many established medical treatments also arise from traditional medical practices. For example the use of aspirin as an analgesic (pain killer) has its roots in the use of tree bark for similar purposes throughout history. The difference between established medicines like aspirin, and alternative medicines such as homeopathy, is that the former have been found to be effective when exposed to rigorous scientific trials.
A form of alternative medicine that has recently been subjected to scientific scrutiny is the use of magnetic bracelets as a method of analgesia. It effective, such therapies would provide cheap and easy-to-implement treatments for chronic pain such as that experienced in arthritis. Unfortunately there is little evidence of such treatments being effective. A meta-analysis of randomised clinical trials looking at the use of magnet therapy to relieve pain found that there was no statistically significant benefit to wearing magnetic bracelets (1). However it can be argued that existing clinical trials may have been hampered by the difficulty in finding a suitable control condition.
The placebo effect
The ‘placebo effect’ is a broad term used to capture the influence that knowledge concerning an experimental manipulation might have on outcome measures. Consider a situation where you are trying to assess the effectiveness of a drug. To do this you might give the drug to a group of patients and compare their subsequent symptomatology to a control group of patients who do not get the drug. However even if the drug group show an improvement in symptoms compared to the control group, you cannot be certain whether this improvement is due to the chemical effects of the drug. This is because the psychological effects of knowing you are receiving a treatment may produce a beneficial effect on reported symptoms which would be absent from the control group. The solution to this problem is to give the control group an intervention that resembles the experimental treatment (i.e. a sugar pill instead of the actual drug). This ensures that both groups are exposed to the same treatment procedure, and therefore should experience the same psychological effects. Indeed this control treatment is often referred to as a ‘placebo’ because it is designed to control the placebo effect. The drug must exhibit an effect over and above the placebo treatment in order to be considered beneficial.
A requirement for any study wishing to control for the placebo effect is that the participants must be ‘blind’ (i.e. unaware) as to which intervention (treatment or placebo) they are getting. If the participant is aware that they are getting an ineffective placebo treatment, the positive psychological benefits of expecting an improvement in symptoms is likely to disappear, and thus the placebo won’t genuinely control for the psychological effects of receiving an intervention.
A placebo for magnetic bracelets
The obvious placebo for a magnetic bracelet is an otherwise identical non-magnetic bracelet. However the problem with using non-magnetic bracelets as a control is that it is easy for the participant to identify which intervention they are getting, as it is easy to distinguish magnetic or non-magnetic materials. The can be illustrated by considering a clinical trial which appeared to show that magnetic bracelets produce a significant pain relief effect (2). In this study participants wore either a standard magnetic bracelet, a much weaker magnetic bracelet or a non-magnetic (steel) bracelet. The standard magnetic bracelet was only found to reduce pain when compared to the non-magnetic bracelet. However the researchers also found evidence that participants wearing the non-magnetic bracelet became aware that it was non-magnetic, and therefore could infer that they were participating in a control condition. This suggests that the difference between conditions might be due to a placebo effect, as the participants weren’t blind to the experimental manipulation.
This failure of blinding was not present for the other control condition (weak magnetic bracelet) presumably because these bracelets were somewhat magnetic. As no statistically significant difference was found between the standard and weak magnetic bracelets it could therefore be concluded that the magnetic bracelets have no analgesic effect. However it could also be argued that if magnetism does reduce pain, the weaker bracelet may have provided a small beneficial effect which might have served to ‘cancel out’ the effect of the standard magnetic bracelet. The study could therefore be considered inconclusive as neither of the control conditions were capable of isolating the effect of magnetism.
More recent research
Recent clinical trials conducted by researchers at York University has tried to solve the issue of finding a suitable control condition for magnetic bracelets. Stewart Richmond and colleagues (3) included a condition where participants wore copper bracelets, in addition to the three conditions used in previous research, while researching the effect of such bracelets on the symptoms of Osteoarthritis . As copper is non-magnetic it can act as a control in testing the hypothesis that magnetic metals relieve pain. However as copper is also an traditional treatment for pain, it does not have the drawback of the non-metallic bracelet regarding the expectation of success. The participant is likely to have the same expectation of a copper bracelet working as they would for a magnetic bracelet.
The study found that there was no significant difference between any of the bracelets on most of the measures of pain, stiffness and physical function. However the standard magnetic bracelet did perform better than the various controls on one sub-scale of one of the 3 measures of pain taken. However this isolated positive effect was considered likely to be spurious because of the number of comparisons relating to changes in pain that were performed during the study (see 4). The same group has recently published an almost identical study relating to the pain reported by individuals suffering from Rheumatoid Arthritis rather than Osteoarthritis (5). Using measures of pain, physical function and inflammation they again found no significant differences in effect between the four different bracelet types.
The existing research literature seems to suggest that magnetic bracelets have no analgesic effect over and above a placebo effect. The use of a copper bracelet overcomes some of the problems of finding a suitable control condition to compare magnetic bracelets against. One argument against using copper bracelets as a control is that as they themselves are sometimes considered an ‘alternative’ treatment for pain, they may also have an analgesic effect. Such an effect could potentially cancel out any analgesic effect of the magnetic bracelets when statistical comparisons are performed. However copper bracelets did not perform any better than the non-magnetic steel bracelets in either study (3, 5) despite the potential additional placebo effect that might apply during the copper bracelets condition. Indeed on many of the measures of pain the copper bracelet actually performed worse than the non-magnetic bracelet. The copper bracelet can therefore be considered a reasonable placebo to use in research testing the analgesic effect of magnetic bracelets.
Despite the negative results of clinical trials, it may be wise not to entirely rule out a potential analgesic effects of magnetic bracelets. Across all three studies (2, 3, 5) the measures of pain were generally lowest in the standard magnetic bracelet group. Indeed significant effects were found in two of the studies (2, 3) although these were confounded by the aforementioned problems concerning control conditions and multiple comparisons. Nevertheless it could be argued that, given the existing data, magnetic bracelets may have a small positive effect, but that this effect is not large or consistent enough to produce a statistically significant difference in clinical trials. This theory could be tested by conducting trials with far more patients (and thus greater statistical power) or by using a number of different bracelets of differing magnetic strengths to see if any reported analgesic effect increases with the strength of the magnetic field. Until such research is performed it is best to assume that magnetic bracelets do not have any clinical relevant analgesic effect.
Image courtesy of FreeDigitalPhotos.net
(1) Pittler MH, Brown EM, Ernst E. (2007) Static magnets for reducing pain: systematic review and meta-analysis of randomized trials. CMAJ 2007;177(7):736—42.
(2) Harlow T, Greaves C, White A, Brown L, Hart A, Ernst E. (2004) Randomised controlled trial of magnetic bracelets for relieving pain in osteoarthritis of the hip and knee. BMJ 329(7480):1450—4.
(3) Richmond SJ, Brown SR, Campion PD, Porter AJL, Klaber Moffett JA, et al. (2009) Therapeutic effects of magnetic and copper bracelets in osteoarthritis: a randomised placebo-controlled crossover trial. Complement Ther Med 17(5–6): 249–56.
(5) Richmond SJ, Gunadasa S, Bland M, MacPherson H (2013) Copper Bracelets and Magnetic Wrist Straps for Rheumatoid Arthritis – Analgesic and Anti-Inflammatory Effects: A Randomised Double-Blind Placebo Controlled Crossover Trial. PLoS ONE 8(9):