Citation check for “exercise is good for chronic pain”
Look at it one way, and it does seem like exercise is an effective treatment for pain. Look at it another, not so much. The general explanation for this inconsistency is probably that it depends on exactly what we mean by “exercise” and by “pain.”
And by “treatment.” Maybe even “for”!
But it gets complicated fast as you dig into the specifics. I want to untangle that paradox someday, and I am working on it — one of my many back-burner projects, but I’m almost ready to move it to the front burner.
Meanwhile, today’s post is just one piece of the puzzle. A study…
| title | Physical activity should be the primary intervention for individuals living with chronic pain A position paper from the European Pain Federation (EFIC) 'On the Move' Task Force |
|---|---|
| journal | European Journal of Pain |
| Volume 28, Number 8, Sep 2024, 1249–1256 | |
| authors | Henrik Bjarke Vaegter, Marja Kinnunen, Jonas Verbrugghe, Caitriona Cunningham, Mira Meeus, Susan Armijo-Olivo, Thomas Bandholm, Brona M Fullen, Harriet Wittink, Bart Morlion, and Michiel F Reneman |
| links | publisher • PubMed • PainSci bibliography |
Vaegter et al argued that exercise is quite a good treatment for chronic pain. Refreshingly they acknowledged — in the first sentence, no less — that “physical activity is challenging for many individuals living with chronic pain.” Their paper is based on the opinion that (emphasis mine) “physical activity should be the primary intervention for individuals living with chronic pain,” because “evidence supports the efficacy of PA in alleviating pain, both in the short and long term, for a considerable portion of individuals with many different chronic pain conditions.”
Does it though? How do they support that statement?
They give five citations from 2017 to 2021. Let’s take a closer look at them, shall we? Do those five critical citations actually back up the premise of the paper? Do they show that exercise alleviates pain, for a “considerable portion” of patients? I applied my citation quality scoring system to each of these:
- Strong support
- Okay support
- Weak support
- No support
- Negative support! A citation that actually undermines the claim.
And here are the citations and their scores, judged on how well much support they seem to provide based on their conclusions, and assuming that those conclusions are in turn actually supported by their data (which is hardly guaranteed either, but I only wanted to go so far down the rabbit hole today):
- Zhang et al — A systematic review and expert consensus with a firm positive conclusion that “exercise is helpful to alleviate neuropathic pain intensity.” 4 points!
- Roos et al — A study of a well-known education and exercise program (GLA:D) for hip and knee osteoarthritis, showing “43-47 % of the patients experienced clinically relevant pain reductions.” That is indeed a happy result, but we also wish it was stronger. The downhill slide has begun: 3 points.
- Vaetger & Jones — This one is a self-cite, the first author citing his own past work. That disqualifies it from getting top marks … but it’s worse than that. It’s an exhaustive review of evidence and possible mechanisms of the short-term effects of exercise on many kinds of pain. They report evidence of some relief with some kinds of chronic pain, but also plenty of evidence that it can backfire with many other kinds of chronic pain. This citation does not obviously endorse the view that exercise “alleviates” chronic pain for a “substantial portion” of patients. What it does show is that it’s messy and unclear, especially for chronic pain. It’s not nothing, but there is only enough good news here for 2 points.
- Franco et al — Another review, this time of fifty studies of exercise for nociplastic (“weird”) chronic pain, reporting effects that were “generally below the minimal clinically important difference threshold,” based on “very low- to low-quality of evidence, creating a high degree of uncertainty.” They recommend exercise anyway, because the lousy evidence does show some better-than-nothing benefits. But the actual results were arguably a negative, and therefore … 1 point, “and you’ll thank me for it” (Greg Davies).
- Geneen et al — A huge review of reviews, crunching about as much data as we ever see in the science of pain, and concluding that … the quality of evidence was low at that time (not that long ago), showing inconsistent and mediocre results at best. This is not support. 0 points.
4, 3, 2, 1, 0: such a tidy batch of scores! With a disappointing total. Out of a maximum possible quality score of 20, these earned only 10 points! I’m no arithmetician, but I think that's 50%, and so that is an “F” for a fail, folks. I’ve definitely seen worse — lower scores are routine with junkier science, even falling to zero. But 50% is very much not great.
Maybe just-right exercise is helpful for certain kinds of chronic pain, and maybe future science will clearly show that. But these five sciences? Not really even close! Boo.