It's an interesting study.
One of the main outcomes is the McGill Pain Questionnaire, probably the most widely used measure of pain, but they don't report the data from this measure. Instead they report the results from a numeric rating. Also well validated. But, why not report both? Did the MPQ scores not behave like the NRS? Odd.
Second, their dichotomization into "relieved" and "not relieved" is not consistent with the majority of the pain literature. Ordinarily, we use a criterion of at least 50% pain relief over baseline. This shouldn't change the results, so I'm not sure why they didn't adopt it.
If I understand correctly, they apply the magnet and the patients may do as they wish for 45 minutes (first, I would have controlled what they did, but you'd expect activity to be randomly distributed between the two groups). Nonetheless, I am in a study where I have a 50% chance of having a magnet strapped to me (they don't tell us anatomically where it is). I am alone. Would I not test to see if this disc/card/strip was magnetic? In my experience, patients are constantly trying to guess if they received the active or placebo intervention. The authors needed to check the validity of their blinding.
Just some thoughts before my Sunday morning coffee...
