Plentiful potent PEMF prevails (defying my cynical prediction)

Late last year I sent out my doubtful first impression of a study of pulsed electromagnetic field therapy (PEMF) for shoulder pain. But my cynicism was not justified in this case!

The results of Kandemir et al seemed too good to be true based on the scanty detail in the abstract alone:

The results that Kandemir et al. report were that PEMF was “superior” to a fake. But is this true and meaningful, or only technically true? (Or any kind of true?) We don’t know how superior PEMF was, because the abstract in this new paper isn’t very chatty, and I don’t have the full text yet. That brevity could be hiding many sins. The authors don’t report how large the benefit of genuine PEMF was (the “effect size”). And who would hold that number back if it’s genuinely impressive?

But you don’t pay me the big bucks to point and laugh at just abstracts, and so I promised an update when I got the full text — and I fully expected to be disappointed.

This is a very common disappointment in my job.

An amazingly positive update

I finally got around to looking at the full text, and … it’s … good?

You could knock me over with a feather! The paper is a clear description of a sensible experiment with properly positive results. Not insanely good, not unbelievably good — but the kind of good you don’t have to squint to see. PEMF really seemed to help.

One study cannot prove that PEMF is actually effective — especially when other studies contradict it. There could be hidden flaws — that is always possible, and statistical jiggery pokery can be really subtle. But there is no obvious problem with it, and so I am pleasantly amazed!

This just does not happen, and it’s a rare pleasure to be reminded that it’s possible.

Kandemir et al. is an exception that proves the rule that most papers aren’t nearly as good as the abstract makes them sound.

The dosage question: it’s really an awful lot of PEMF’ing

There’s got to be something wrong this good news … and the fly in the ointment is probably the dosage. The dosage may explain why it worked, but it also explains why it will probably never work for you.

My last post reported on a similar kind of study of shockwave therapy (Heide et al.). Both compared a real medical gadget to a fake one, but the shockwave one had the cynically expected results. But were the shockwaves shocking enough? A weakness of that study was testing such a low dosage, which some readers griped about.

Testing a modest dosage does not render a study useless: there are good reasons to test both idealistic and more realistic treatment protocols. Of course we’d like to know if a treatment works at any dosage, even if it’s too much to be useful in practice.

But we also want to know if a treatment can deliver any happiness that normal people can actually afford. Most of us won’t keep trying a treatment that doesn’t show a benefit within about three hours or three hundred bucks. The point of trialling an affordable dosage isn’t to show that it can finish the job, but to show it can start the job.

The PEMF study went the idealistic route, testing truly plentiful PEMF: 20 half-hour sessions, daily for 5 days a week for 4 weeks. And it was potent PEMF, too! Kandemir et al reckon that’s why it worked, while smaller dosages failed in other studies.

But … daily half-hour appointments for a month? That’s too expensive for most of us in time alone, never mind dollars — few people can afford to spend that much on a rehab boost that is merely “good” and not “great” (and even good is hardly guaranteed, because nothing works for everyone). So you tell me: does PEMF really “work”? Even if the result is true, is it actually good news? Would you do 20 half-hour sessions of PEMF for probably “good” results?

Both PEMF and shockwave therapy could have a good therapeutic effect at impractical dosages … but also fail at more affordable dosages.

•

My full guides to both therapeutic ultrasound and electrical stimulation treatments have been updated with these citations. Those are both fully free articles in the main PainSci library.