A slightly edited excerpt from Are Electromagnetic Fields Making Me Ill? by Brad Roth
In biology, much of the discussion about how magnetic fields interact with the body focuses on mechanisms: how does a response happen. This mechanistic approach is usually taken by physicists and engineers. But most medical doctors don’t care how magnetic fields interact with the body, but only if magnetic fields interact with it. In particular, they’re concerned mainly about a magnet’s usefulness in the clinic. For example, do static magnetic fields alleviate pain? To answer these types of questions, we need clinical trials.
The problem is, many clinical trials have been conducted to test the hypothesis that magnetic fields reduce pain, and the results are inconsistent. What we require is a procedure to weigh the pros and cons of all the clinical trials, placing greatest weight on the best ones. What we need is meta-analysis.
Meta-analysis is a formal, quantitative method of systematically assessing a group of clinical trials, to reach conclusions that are more reliable than those of any individual study. In 2007, Max Pittler, Elizabeth Brown and Edzard Ernst published “Static Magnets for Reducing Pain: Systematic Review and Meta-Analysis of Randomized Trials” (Canadian Medical Association Journal, Volume 177, Pages 736–742). They analyzed 29 relevant clinical trials. The best of them were placebo-controlled, randomized, and double-blind.
All the trials were placebo-controlled. This means that some patients received treatment with real magnets, and others were treated with objects that resembled magnets but produced a much weaker magnetic field or no magnetic field at all (the placebo). Doctors know that sometimes a patient responds positively if they think they’re being treated, regardless of whether or not they receive the treatment. This is a problem, especially for studies that assess subjective judgement, such as the severity of pain. In a placebo-controlled experiment, the patient can’t determine if they’re receiving a treatment or merely a well-designed placebo. Whenever possible, clinical trials should be placebo-controlled.
All the trials were randomized. That means that patients were placed in the treatment or placebo groups at random, which avoids spurious correlations. For example, you don’t want patients in the placebo group to consist of more men, while the treatment group is more women, since then you would have difficulty distinguishing treatment effects from gender effects. Similarly, you don’t want the treatment group to consist of young patients with the placebo group having old patients, or poor patients compared to rich patients, or fat patients compared to thin patients. By assigning patients to the placebo or treatment group at random, you expect on average equal numbers of men and women in each group, equal numbers of young and old, rich and poor, fat and thin. This averaging works best in large studies, where random fluctuations are less likely.
Finally, most of the trials were double-blind. Obviously, you don’t want patients to know if they’re in the placebo or treatment group, otherwise the placebo effect may not work its magic. That is, you want a blind study, with patients not aware whether or not they received the actual treatment. Double-blinding means that not only the patient but also the physician doesn’t know who’s in the placebo or treatment group. This eliminates any subtle hints that a physician might let slip when delivering the treatment. Try as they might to treat each patient the same, it’s difficult for doctors to avoid providing inadvertent clues if they know which group a patient is in.
The finest clinical trials are therefore placebo-controlled, randomized, double-blind, and have a large number of patients in the study to avoid statistical fluctuations. A meta-analysis takes all these factors and others into account when coming to conclusions.
In their meta-analysis, after weighing the evidence from all the clinical trials, what did Pittler, Brown, and Ernst conclude? They wrote “The evidence does not support the use of static magnets for pain relief, and therefore magnets cannot be recommended as an effective treatment. For osteoarthritis, the evidence is insufficient to exclude a clinically important benefit, which creates an opportunity for further investigation.” In other words, in most cases the evidence goes against magnets providing pain relief. In the case of arthritis, there’s not enough evidence to say one way or the other.