Stretching for fibromyalgia: abstract versus full text
On the heels of another round of my stretch “negativity,” today I’m reporting on a new study (Støve et al.) that challenges my notorious anti-stretching bias, and offers fibromyalgia patients a smidgen of hope.
The study reports that a mere six minutes of daily home stretching for six weeks reduced fibromyalgia symptoms by a little over 18%. Not exactly “huge” if true, but it’s good, and cheap, and no major flaws leap out of the abstract, so … should I put that in my pipe and smoke it?
Not today. [YouTube, Game of Thrones scene] Based on the abstract alone, it sure looks like solid good news, but abstracts are like movie trailers: they only show you the good bits.
The honest conclusion — like an “honest trailer” for a bad movie [YouTube again] — is that a six-minute daily routine with weekly researcher contact improved self-reported fibromyalgia symptoms over six weeks compared to — wait for it — being on a waitlist. So the “result” is hardly a surprise, and the reality is a much weaker claim than “stretching works for fibromyalgia.” My bias is safe from this junk.
The critical flaw here is a cliché of amateur science: a failure to adequately “control” major confounders, AKA “the whole point of doing a clinical trial.” The paper title claims it is “controlled,” but it’s just not: the “usual care” waitlist design is a what-were-they-thinking choice.
This trial cannot separate the effect of stretching from the effects of expectancy, novelty, and weekly contact with a caring researcher, yada yada yada — an especially glaring problem with fibromyalgia, where patients are starved for validation. Participants weren’t blinded either, so the waitlisted group results were probably deflated by disappointment (“frustrebo” effect, performance bias), widening the artifactual gap further. All of this has happened before, like with massage for fibromyalgia.
The authors acknowledge this problem, but they seriously understate how decisively it sabotages the experiment:
“the waitlist design may have induced expectancy asymmetry and limited causal inference compared with an active comparator”
This is like saying that eating potato salad left in the sun for an afternoon may induce facilitated fecal transit. The waitlist design definitely broke the experiment.
I’ve recently reported that some people with fibromyalgia do not think stretching even feels good. This paper’s conclusions are implausibly inflated in favour of stretching, and if anything the study just adds to the pile of accidentally produced evidence that fibromyalgia benefits more from compassion, validation, and any light activity than any other known therapy.
Which is why I will continue my easy daily stretching routine — ideally with a dog.
Dogs are a superb source of the validation and compassion that were probably the active ingredients in the Støve et al. experiment. A good dog will provide these while you stretch (sometimes while napping).