Clinical effectiveness of an online group physical & mental health rehab programme for post-covid-19 condition REGAIN study, 2024, McGregor+

££££££

 

Since the study's definition of full adherence was nothing of the kind & partial adherence could just be attending the introductory session, this is deeply unimpressive. If participation was genuinely voluntary, participants must have hoped this would work for them and be committed to it. Therefore this really didn't work for participants or they were too sick to participate, which suggests a lot more PEM/PESE than the researchers have admitted to.

The NHS was always going to offer this anyway and it probably mirrors standard PIC, cardiac & pulmonary rehab, for which there must be plenty of evidence. The only question for them, really, was whether more affordable remote sessions were as acceptable as face-to-face. They have no interest in or knowledge of PEM/PESE or concern for this kind of harm. Why not just skip this expensive trial and ask patients what mode they prefer or offer a mixture of in-person and remote?

Or is the trial a dog whistle to elicit these headlines and tweets from the usual suspects?

 

The difference for the primary outcome was 0.03 on the PROPr questionnaire which apparently has a scale from -0.22 to 1. Table 2 shows that the standard deviation is somewhere around 0.18 so the effect size is approximately 0.16 of the standard deviation, which is very, very small.

"A difference of 0.03 to 0.05 is considered to be clinically important" the authors write but when I looked at the reference they provided (reference 13), it says":

What it this minimally important difference for PROPr?
The minimally important difference for PROPr has not been formally evaluated, but we currently recommend using 0.04. Most preference-based measures have minimally important difference thresholds between 0.03 and 0.05. Current work-in-progress suggests a minimally important difference of 0.04 would be appropriate, although a conservative estimate of 0.08 (half of a standard deviation in the PROPr dataset) could also be used.​

https://www.proprscore.com/faqs/

So if the authors had followed this guideline (don't know if it is any good) they should have said that the effect they found probably isn't clinically relevant.

 

Yes I was really struck by how minuscule that difference was. Even if it were barely above the minimally significant threshold, it’s still would be miniscule.

 

That this is being touted as impressive says a lot. It could not be a smaller effect, could not be more biased or misleading in having excluded the main problem in Long Covid and chosen the least likely cohort of hospitalized patients, while most with LC were not. And yet it's IMPRESSIVE in bold lettering.

And of course this is not the first trial of this. Most report pretty the same very underwhelming conclusions, the bare minimum of uncontrolled indirect benefits with an open label unblinded outcome that doesn't even report physical benefits out of a physical training program. With every study that reports meagre benefits, it's like the slate is wiped clean everything, but at the same time it confirms other studies. No, not those studies. Those other studies, most of which conclude that more evidence is needed, because it's basically the whole game.

It's especially frustrating seeing just how mediocre this all is, and yet it has come to completely dominate and ruin millions of lives while being less impressive than the stuff that comes out of a dog's butt.

 

From @MSEsperanza:

Science Media Centre (UK) did a briefing that was accompanied by an SMC Roundup of comments:

https://www.sciencemediacentre.org/...-with-long-covid-after-covid-hospitalisation/

Seems strange to me how two of the commenters can mention some crucial criticism (much in accordance with forum members') but at the same time praise the study.

E.g.: Dr Baptiste Leurent, Associate Professor in Medical Statistics, UCL, said:

“A key limitation of such study is that participants are aware of the intervention they receive, and could well score more positively questionnaires if they enjoy the programme. Without objective measures, it is difficult to disentangle the actual benefit of the programme in addressing long covid symptoms, from the positive perception of receiving such programme."

 

Good catch! They're really misrepresenting the guidance in a way that makes their pathetic findings look marginally better. Also ironic is their claim that the study could be under-representing the effect--if they had compared it to no treatment at all instead of the standard care. It's like they want to maximize their lack of having controlled for all the attention. So they're pitching these as "conservative" findings in a way.

 

I don't know much about minimally clinically important differences but the idea looks confused to me.

If you have a scale from 0 to 1 in roughly 0.1 steps then I would expect the minimum change relevant to a patient to be 0.2, just possibly 0.15.

But in a trial where you have a large enough sample to be sure of the size of effect (which here seems barely the case f at all) then you would be happy to show that a third of subjects achieved a minimally important difference. So the minimally important difference for a trial population could be 0.05.

I suspect they built this in to their calculations but it is not clear from much of the literature.

For something like preventing stroke or cancer - with something like a statin - I can see that quite small effects would be worth considering. However, in this sort of context I would be unimpressed by a difference for a trial population of less than 0.8 think. And that would require a design with minimum risk of bias.

So this goes less than nowhere.

 

Just noticed that in their discussion section they try to justify their interpretation as follows:

"Research completed since we started this study suggests a minimally important difference of 0.04 on the PROPr score between groups. Our observed differences of 0.03 (95% confidence interval 0.01 to 0.05) at three months and 0.03 (0.01 to 0.06) at 12 months are smaller than this suggestion. However, the complier average causal effect analysis showed a larger effect of 0.05 (0.01 to 0.09) at three months and 0.06 (0.01 to 0.10) at 12 months, suggesting that the true effect, in those fully complying with the intervention, might exceed this threshold. Our post hoc analysis showing numbers needed to treat of 11.9 for “much better now” and 5.4 for “much better now” or “somewhat better now” combined at three months will help to interpret the clinical importance of our findings."
​

 

They should just report the effect size, which will likely be around than 0.15 standard deviations which is very small. Hard to take this seriously in a trial that did not control for attention, expectations etc.

 

Think Jonathan's highlighted before that the NHS needs more money & I think it's underfunded compared to e.g. Germany & France. Also, that he (Jonathan) retired when they wouldn't come up with the resources to do the job.
Looks like this may be an example of an alterative strategy* --- makes it appear you're doing something --- OK you're not doing anything significant --- who cares (other words come to mind) --- not my family members---.

*"Clinical effectiveness of an online supervised group physical and mental health rehabilitation programme for adults with post-covid-19 condition (REGAIN study): multicentre randomised controlled trial"

 

Published in the BMJ? Is there an emoji for toilet paper?

 

Remember the add "because you're worth it" well, this is the opposite. Alternately - because they couldn't care enough to do a good job -- or other words!

 

I was really stunned seeing this. It's so brazen. They're basically arguing that less rigorous is better, as has been increasingly argued in recent years, that you can simply imagine it to be better and pretend that it could be. It makes zero sense as an argument, but that's just par for the course with biopsychosocial trials.

 

Copied from the SMC news thread.

There is "Expert reaction" here:

https://www.sciencemediacentre.org/...-with-long-covid-after-covid-hospitalisation/


to the BMJ paper:

‘Clinical effectiveness of an online supervised group physical and mental health rehabilitation programme for adults with post-covid-19 condition (REGAIN study): multicentre randomised controlled trial’ by Gordon McGregor et al. published in the BMJ at 23:30 UK time on Wednesday 7 February 2024.

DOI: 10.1136/bmj‑2023‑076506

https://www.sciencemediacentre.org/...-with-long-covid-after-covid-hospitalisation/

 

Indeed, and that seems to have been picked up in a whole paragraph of one of the expert commentaries regarding the drop-out rate basically hiding the real results.

I also suspect they in effect drove the drop-out rate by making it clear what their attitude to 'PEM' was, according to the statnews article ( Long Covid study tests exercise as treatment option (statnews.com) :

"Fatigue leads the list of persistent problems experienced by people with long Covid — which is why patients have pushed back against treatment approaches that endorse escalating levels of exercise for a condition that researchers are still trying to understand. They fear post-exertional malaise, the debilitating price to be paid for pushing their bodies too hard.

Researchers from the U.K. were well aware of those potential harms. Working with patients, they designed a trial called REGAIN to test a way to provide mental health therapy and exercise guidance to people with long Covid without making their conditions worse. Their goal was to help people manage their symptoms, improve their functioning, and reduce their distress."

"Early on, the researchers were contacted by members of advocacy groups who expressed concerns about exercise training.The fear is that, as in another post-viral condition, myalgic encephalomyelitis or chronic fatigue syndrome, intense fatigue and worsening of other health issues after physical or mental activity will follow exertion. A clinical trial looking at exercise in RECOVER, the controversial long Covid research project from the National Institutes of Health, prompted backlash when an exercise study was announced in late 2022."


I can't fully tell whether they just basically saw PEM as 'fear of PEM' so Chalder-esque ie that they emperor's new clothes that it doesn't exist and anyone saying anything like it as a symptom or reaction got accused of 'having the fear'? Or whether they believe they 'cracked it' with the 'it's the way it was done by the last lot' type thing.

I am however, a bit concerned by this paragraph mentioning a companion editorial, and whether it understood what such a high drop-out rate could mean vs the conclusions it is claiming to draw:

"A companion editorial approved of the trial’s approach: “Regular monitoring in the REGAIN trial did not identify any episodes of post-exertional exacerbations of symptoms, providing reassurance that individualized exercise at home in online groups supervised by a trained physiotherapist or exercise physiologist is safe.”"

 

From: Rehabilitation for post-covid-19 condition | The BMJ :

"One serious adverse event (syncope and vomiting after exercise) was deemed possibly related to the intervention, and two adverse events were deemed definitely related (knee pain with exercise, and severe anxiety before exercise)."

Who does the 'deeming' on these assessments of harm and 'whether it was related'?

 

A rather poor New Scientist piece:
https://www.newscientist.com/articl...ps-people-with-long-covid-but-its-no-panacea/

Concluding
On the other hand, it is the first time a randomised trial – the best medical evidence – has shown anything to have even a modest benefit for this condition. That should surely be a cause for cautious celebration.

Despite acknowledging this is a controversial area, they only reference MEcfs campaigners etc, no mention of the scientific critiques; not a word on the problem of subjective outcomes in unblinded trials