Discussion in 'PsychoSocial ME/CFS Research' started by Cheshire, Jan 4, 2021.
The only randomized trial for patients diagnosed with criteria where PEM is required is the trial by Pinxsterhuis et al. 2015. It tested a self-management program based on the Energy Envelope Theory for ME/CFS patients diagnosed with the Canadian criteria.
The intervention did not have any sustained effect compared with usual care.
They classify definitions as chronic fatigue (no PEM), CFS (PEM optional) and ME (PEM and other symptoms required). Oxford was classified as a chronic fatigue definition, the London criteria as a CFS criteria and the rest as you might expect although the interpretation that NICE requires PEM is incorrect IMO (NICE considers PEM's symptom exacerbation to be optional)
But they rate study methodology much higher than would appear to be warranted - 15 of 19 studies rated high quality including PACE, 4 as fair quality.
Yes, although their hearts may be in the right place I get the impression that they have no idea of the weakness of virtually all physio trials. Perhaps not surprisingly.
I am not sure that this emphasis on diagnostic categories helps. I think it probably distracts from the genuine scientific problem - that the methodology is no use.
I agree that use of bad methodology like unblinded studies with subjective outcomes is a genuine scientific problem.
But selecting the wrong patient population is also a genuine scientific problem. Making claims of efficacy and safety for a patient population based on studies that included patients who did not have the disease is a real problem and has caused decades of harm to patients.
And when you have both of these at the same time, it only magnifies the extent of the scientific problem. We need to fix both problems
If they included PACE and FINE in this, nevermind rating those as high quality, they have a very loose definition of physiotherapy that is far past the point of irrelevance. Is physical mixed with cognitive therapy run by nurses or clinical psychologists really physiotherapy? They aren't even controlled so, again, misuse of common words.
Honestly nothing useful here, definitions and criteria all over the place and no consistency. This doesn't tell us anything about what role physiotherapy could play. Which I assume is little, but this study tells us nothing about it.
Not read the paper yet, but just emphasising how little we know in the abstract is still of some value.
It may be a practical shortcoming but it is not actually a scientific problem. In all science you have to cast your net wider than the conditions that define any individual. If you cast wide your predictions can reasonably be applied to other cases that fall under the same classification. What is less legitimate is to cast your net narrowly and then extrapolate.
I think the original criticism of PACE on this basis was wrong. The PACE authors were perfectly entitled to study a broad group and produce evidence relating to that broad group. It would have been reasonable to be cautious about subgroups with adverse reactions to exertion but in reality this was completely overshadowed by the more important shortcomings of the trial. The CDC or IOM (?) completely misread this in my view. As someone who has designed and executed trials I could not fault PACE on that basis scientifically.
"A self-management program including eight biweekly meetings of 2.5 hours duration. The control group received usual care."
Biweekly, so twice a week or every two weeks?
If twice a week I can see this treatment easily being counterproductive on the basis of the exertion alone.
I don't have a problem with pacing not being better than placebo but it should be tested properly. The pacing developed by patients is so simple that it doesn't require much time to learn or any therapist input. It's also something that patients will figure out on their own so pacing education in patients that have been ill for a while might add no benefit, not because it doesn't work but because the patient is already getting any benefits it can provide.
I think we'll just have to agree to disagree
Its difficult for me to imagine that it would be scientifically legitimate to cast the research net so wide that it would ignore whether the key feature of the disease is present or not - particularly when that feature would actually predict harm from the treatment being studied. At that point, the ME cases do not "fall under the same classification." IMO, it's not appropriate to view ME as cases or a subset under the classification of Oxford CFS. Lots of conditions are associated with chronic fatigue. Why is ME the only one considered a subgroup of Oxford CFS for which Oxford-based efficacy and safety claims be legitimately applied?
Re CDC and IOM - IOM didn't include Oxford or its studies but AHRQ and the NIH report both recommended Oxford not be used because it could include patients with other fatiguing conditions. What the IOM did do is discredit the claim that childhood trauma/abuse put people at a six fold risk of developing CFS. Their rationale was that the supporting study used the Reeves definition which was so broad that it could include an overrepresentation of PTSD and depression.
As AHRQ and NIH note, a similar possibility exists with PACE. I agree that study conduct issues are damning. But even if study conduct were perfect, could you really legitimately generalize the efficacy and safety findings of an Oxford study to people with ME? Both need to be considered.
i have never been told by a doctor to pace or even to take things easy since january 1990 not once was i given any advice it was a case of here's your diagnosis now lets forget it.
A systematic review of randomized controlled trials published over the last two decades was conducted.
Shouldn't have taken long.
I think they focus on diagnostic categories because this looks to be a review aimed at physiotherapists wanting to make practical treatment decisions based on evidence. It's very reasonable to suspect that people with and without PEM may react differently to some treatments and therefore to look for evidence specific to people with and without PEM, rather than extrapolate from one to the other.
Unfortunately that's a rather difficult project when the evidence base is as poor and all over the place as what they had to work with here, and they do point out some of the main methodological issues, too.
Their findings are no surprise to us but it's good to see it clearly stated:
that methodological problems are significant and explain most of the "positive" findings even in cohorts which include people without PEM,
that almost nobody has looked at any sort of physiotherapy treatment in cohorts which include only people with PEM,
that harm reporting is woefully inadequate, and
that based on this poor evidence physios should not offer any treatment beyond "symptom relief and increasing or maintaining health status and quality of life, by improving the ability to cope with ME, guiding self-management and avoiding PEM".
Agreed on the extrapolating from a narrowly defined group, not legit. But surely this applies the other way round, too. You can't study a group selected by very broad criteria and then claim the overall findings apply to all subgroups - unless you've actually properly looked at the subgroups. That would be like selecting a group of headache sufferers, mostly of the common-garden type with maybe or maybe not a small handful of migraineurs mixed in, give them all an aspirin, find it helps a good number, and then claim aspirin treats migraine.
Yes. This looks to be the result of using the PEDro scale to assess methodological quality. I don't know if there are better tools out there they could have chosen? At any rate the PEDro scale by itself is clearly not fit for purpose here. In fairness, further down in the article the authors mention a few of the methodological clangers committed in PACE and other studies and they particularly stress the problems with relying almost solely on subjective outcomes in these unblinded trials. But this could have been done better.
Intervention effects, if any, seemed to disappear with more narrow case definitions, increasing objectivity of the outcome measures and longer follow-up.
symptom relief and increasing or maintaining health status and quality of life, by improving the ability to cope with ME, guiding self-management and avoiding PEM.
I don't feel I have a good understanding of PACE, but I'm slightly surprised by this "As someone who has designed and executed trials I could not fault PACE on that basis scientifically."
The reason I say that is that those involved in PACE seem to have failed to incorporate adequate indicators of whether the intervention worked. E.g. Fluge and Mella used acivity monitors to objectively assess activity levels. PACE used subjective questionnaires. Even those with limited knowledge may have come across the Hawthorne Study [https://en.wikipedia.org/wiki/Hawthorne_effect]. When you study people they are grateful for your help.
As well as being concerned by the PACE researchers, I'm concerned that Government policy advisors bought into this i.e. they accepted that GET and CBT would be effective treatments [NICE guidance]. I assumed that is due to "information asymmetry"; however, I'm amazed that NICE didn't have the expertise to evaluate PACE --- that's what they are appointed to do!
This comment relates specifically to the breadth of the definition for inclusion criteria. There is nothing scientifically wrong with using broad inclusion criteria, along as they are adequately defined. Inclusion criteria do not need to match preconceptions of disease categories. Very often broader or narrower criteria are relevant to particular studies.
Apart from this one respect in which PACE is scientifically fine, the trial was a disaster in almost all other respects. That included a design that was bound to introduce selection bias (against selecting for people with known exertion intolerance) which then made a nonsense of using the inclusion criteria as a basis for interpreting the results.
Again I must emphasise that it is always legitimate to sample broadly and then apply results to subsets but it is NOT legitimate to sample narrowly and the apply to instances outside the narrow sample definition. If you test the reliability of Audi cars you cannot then extrapolate to a Trabant. But if you test reliability of all types of car you can at least give statistical predictions for any car. If no car broke down every day for a week you could predict it is unlikely that another car will.
This principle applies to everything we do in science because there are always subsets that have to be grouped together in order to make any generalisation. Every individual is different and the point of science is to try to generalise to further cases. You simply cannot do that for cases that do not fall under your original definition, even if you may be able to make an inspired guess as to whether they would behave the same.
What worries me continually in relation to this is the continuing idea that trials of exercise may be flawed because they are directed at people who do not have exertion intolerance so would benefit and these are confused with people with exertion intolerance who do not benefit. It is time that physios realise that we have no evidence that exercise benefits anyone with fatigue. Continuing emphasis on diagnostic groups distracts from the key message.
I also think there may be a confusion over the intentions of the PACE authors in recruiting. The impression may be given that they were just interested in fatigued people generally and so the trial was not really directed at ME. But I think this is wrong. The theories they had about people being afraid to exercise were clearly directed at people who described payback from exercise. Whether or not that is strictly PEM I think is irrelevant. They had in mind people who would be described as having ME, who have exertion intolerance. They set the inclusion criteria wider maybe, but that is scientifically justifiable, just perhaps expected to be a rather blunt instrument.
If they thought they were including people without exertion intolerance then it would make no sense to use CBT directed at countering exertion intolerance or indeed graded exercise, which is based on an assumption of exertion intolerance. If there was no exertion intolerance then people could just be told to go off and live a normal life.
Is there any evidence that the PACE researchers believed exercise intolerance exists, I had assumed they believed that patients were miss reporting the effects of unfitness or ‘deconditioning’, that all patients were on a single fatigue continuum with no qualitative distinction between those with just the symptom of chronic fatigue and those reporting PEM?
I had assumed they believed the reasons telling people ‘to go off and live a normal life’ did not work because of psychological issues rather than anything to do with exercise intolerance. Obviously they wanted to do things gradually because that is less stressful than shock tactics and is generally preferred in psychological models for treating other avoidance related situations such as phobias, and because getting patients ‘fit’ gradually undermines the ‘false cognitions’ whereas going straight from ‘sofa to running ten miles’ in any unfit person would produce all sorts of negative symptoms that might reinforce the ‘false illness beliefs’.
But that is the point. When they believed patients were 'misreporting' the patients were reporting exercise intolerance. The BPS crowd wanted to study these 'misreporters' - i.e. PWME. I don't think they were interested in people who just said they were fatigued. They were interested in these misreporter people - people who were actually reporting exercise intolerance.
PEM strictly speaking requires delay in payback and I don't think the BPS people were specifically interested in that. But they were interested in people who reported payback. I am unclear whether the delay in PEM is actually a crucial factor. We don't have any particular theoretical reason to think it separates out a different disease mechanism. What I see as the key feature of ME is exertion intolerance in the sense of disproportionate payback. I think the BPS crowd recognised that and theorised (wrongly) that it was central to what their treatments targeted.
Thank you very much for your response - as always --- I may have to think through some parts of it though!
Separate names with a comma.