The evidence base for physiotherapy in myalgic encephalomyelitis/chronic fatigue syndrome when considering PEM: a systematic review (2020) Wormgoor

[rogerblack]

I note the very important and highly misleading line '
Currently, there is no scientific evidence when it comes to effective physiotherapy for ME patients. Applying treatment that seems effective for CF or CFS patients may have adverse consequences for ME patients and should be avoided.'.
“When I use a word,’ Humpty Dumpty said in rather a scornful tone, ‘it means just what I choose it to mean — neither more nor less.’

You don't get to use words like that in the conclusion without pointing out you are using your own definitions for CF, CFS, and ME, as casual readers (clinicians with their own approaches, patients with diagnosis with criteria that will not meet these assumptions) will misread them.

I note again on the topic of PACE that you don't need to dig into anything other than table 4 of the 'cost effectiveness' paper published by the original team, the year after.

Would actiometer/... reports have been nice, sure.
But this right here is perfectly adequate to damn it from any rehabilitation perspective.

 

[Mithriel]

I am not sure I understand the idea of using a wide net. The PACE trial was examining a treatment so any question of whether they were looking at a singe disease or many was surely important especially as they only looked at average outcomes not individual patient results.

If everyone with a diagnosis of "infection" had been used in a trial of an antibiotic then the average could well have been low so it would have been discarded yet it may have been excellent at treating TB.

If the definition is wide then subgrouping of results is essential yet that was not done for PACE. Looking at why some patients are helped and others not has to be part of using a wide definition.

 

[Invisible Woman]

I think I am following @Jonathan Edwards argument about casting a wider net and monitoring subsets of patients in scientific trials, including PACE.

Does this not make the lack of reporting of harms even more egregious though?
Is this not the flip side of casting a wide net? You specifically look for differences between groups, who the treatment is effective for, who it doesn't work on and who it affects adversely.

 

[Medfeb]

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.

So does this mean it would be appropriate to select a cohort based strictly on the basis of shortness of breath to test an asthma treatment and then apply the findings to patients with heart failure and COPD? Perhaps not exactly the right clinical analogy but my point is that inclusion criteria that are TOO broad and thus clinically meaningless will result in meaningless and potentially harmful findings.

And so with PACE. Its singular criterion of medically unexplained chronic fatigue is throwing everyone with no explanation for their condition into a single clinical classification/entity.
But that's a nonsensical entity as discussed elsewhere regarding MUS - an entity based solely on the absence of evidence.

Regarding the PACE authors being interested in studying exercise intolerance... if they had been interested in that, wouldn't they have at least included that as an additional required symptom? But they didn't.

 

[ME/CFS Skeptic]

So does this mean it would be appropriate to select a cohort based strictly on the basis of shortness of breath to test an asthma treatment and then apply the findings to patients with heart failure and COPD? Perhaps not exactly the right clinical analogy but my point is that inclusion criteria that are TOO broad and thus clinically meaningless will result in meaningless and potentially harmful findings.

I suspect that criteria that are too broad will likely result in null results.

As I understand it in most areas of medicine, researchers are usually accused of defining their inclusion criteria too strict; of selecting patients where they think treatment will work. Clinicians than usually reply that this is not a realistic representation of their clinical practice where only a small percentage would meet those criteria, so the study is not as impactful as it claims etc. etc.

So when people argue that the PACE trial has broad inclusion criteria, I think most researchers from outside the field would initially interpret this as a strength, not a flaw.

 

[dave30th]

They were interested in these misreporter people - people who were actually reporting exercise intolerance.

Jo, I find this to be an extraordinarily generous reading of their intentions, especially since I don't believe I have ever read this idea before. Maybe they have said it themselves, but I haven't seen it. I have seen nothing to indicate they were interested in anything other than people who met the Oxford criteria. Also, they went out of the way to blur the distinctions between their broader group and their two sub-groups when making public claims about extrapolating the results.

 

[Invisible Woman]

Regarding the PACE authors being interested in studying exercise intolerance... if they had been interested in that, wouldn't they have at least included that as an additional required symptom? But they didn't.

Also if this was the case wouldn't there have been other important avenues to explore?

If they were studying exercise intolerance, especially in people where deconditioning was a perpetuating factor then why not use healthy, albeit sedentary and deconditioned controls?

If necessary they could have deliberately asked controls to decondition initially.

The PACE trial involved physios so surely it would have been easy enough to recruit and assess a deconditioned, healthy control arm?

 

[Jonathan Edwards]

Jo, I find this to be an extraordinarily generous reading of their intentions, especially since I don't believe I have ever read this idea before. Maybe they have said it themselves, but I haven't seen it. I have seen nothing to indicate they were interested in anything other than people who met the Oxford criteria. Also, they went out of the way to blur the distinctions between their broader group and their two sub-groups when making public claims about extrapolating the results.

There was nothing generous about it. I guess the broad grouping may have been driven by insurance politics as much as lazy mindedness but I don't think one can doubt that the rationale for CBT and GET revolves around the idea of misreading normal payback as pathological payback - which is all about (their own misreading of) ME.

I think the idea that they wanted to study some broad group of people who just happened to be fatigued for no obvious reason does not add up. I am not sure these people exist to be honest.

It is all a muddle. Wessely distinguished 'ME' as people with inappropriate beliefs from 'CFS' which he did not at the time say more about but implied that they were people with genuine fatigue. But the PACE trial does not make sense as a trial for CFS because it was targeted at people with inappropriate beliefs.

 

[Jonathan Edwards]

So does this mean it would be appropriate to select a cohort based strictly on the basis of shortness of breath to test an asthma treatment and then apply the findings to patients with heart failure and COPD?

No, I said that was exactly the wrong thing to do.

that inclusion criteria that are TOO broad and thus clinically meaningless will result in meaningless and potentially harmful findings.

It doesn't work like that. As Michiel says, it dilutes and risks missing a result, but it can be a very good strategy. I used it myself for studying inflammatory arthritis of the knee. It is routinely used for simple painkillers - where diagnosis is largely irrelevant.

And so with PACE. Its singular criterion of medically unexplained chronic fatigue is throwing everyone with no explanation for their condition into a single clinical classification/entity.
But that's a nonsensical entity

No, it is perfectly reasonable to study general categories like this - as for pain. Categories for inclusion in trials do not have to be specific diagnoses. They need to make some sort of sense in terms of the treatment strategy. For PACE that was targeting misinterpreted symptoms of payback. It may be fair to say that the Oxford criteria go too broad for that. But they had a problem. They did not think people were actually exertion intolerant. They thought that patients just thought they were. So including exertion intolerance per se would not make sense. It would be a bit tricky saying to patients that they could only take part in the trial if they were thinking silly thoughts.

I absolutely agree that they hashed everything up in more ways than we have even discovered after arguing about it for five years. But the point remains that broad categories are perfectly sound from a scientific perspective. Even if they have no very good rationale they can provide evidence relevant to a broad category - but not if you then hash up your methodology.

 

[Trish]

So if the PACE researchers had actually listened to people with ME describing the symptoms they get during PEM, and understood that they are very different from the symptoms of tired and aching muscles the next day after unaccustomed exercise in people who are deconditioned, and different again from other forms of chronic unexplained fatigue, they would see that lumping them together doesn't make sense.

As it turned out, CBT/GET didn't work for anyone, so they inadvertently showed that it is useless for all patients with chronic fatigue of any sort, including ME/CFS with PEM.

What if PACE had come out with a clear positive result with a significant number of patients recovering? If, for example, the majority of patients they swept up with their broad brush had depression, deconditioning, fibromyalgia and other forms of fatiguing condition, and a signficant number had improved their activity levels markedly.

Then they would have been able to apparently legitimately claim that CBT/GET works for all the conditions under their broad umbrella Oxford definition. The fact that those with ME/CFS got worse could be swept under the carpet, especially as they didn't do proper recording of harm.

And if they didn't record any differentiation between those with PEM and those with post exertional fatigue through deconditioning, how would they know that pwME were the ones getting worse?

 

[Hoopoe]

So if the PACE researchers had actually listened to people with ME describing the symptoms they get during PEM, and understood that they are very different from the symptoms of tired and aching muscles the next day after unaccustomed exercise in people who are deconditioned, and different again from other forms of chronic unexplained fatigue, they would see that lumping them together doesn't make sense.

There is a third possible explanation which is that they did listen and interpreted what patients were saying as irrational belief and catastrophizing.

 

[Medfeb]

No, it is perfectly reasonable to study general categories like this - as for pain. Categories for inclusion in trials do not have to be specific diagnoses. They need to make some sort of sense in terms of the treatment strategy. For PACE that was targeting misinterpreted symptoms of payback. It may be fair to say that the Oxford criteria go too broad for that. But they had a problem. They did not think people were actually exertion intolerant. They thought that patients just thought they were. So including exertion intolerance per se would not make sense. It would be a bit tricky saying to patients that they could only take part in the trial if they were thinking silly thoughts.

I absolutely agree that they hashed everything up in more ways than we have even discovered after arguing about it for five years. But the point remains that broad categories are perfectly sound from a scientific perspective. Even if they have no very good rationale they can provide evidence relevant to a broad category - but not if you then hash up your methodology.

I agree that inclusion criteria need to make sense in terms of the treatment strategies. I appreciate that a broad strategy across diagnoses might be useful in the case of pain. But even there, wouldn't you need to at least stratify by the type of pain if different medications have differential effects depending on the type of pain. Or potentially design the study to target such the subset of the pain population with that specific type of pain?

But here we are talking about medically unexplained chronic fatigue and the associated unproven, unprovable default assumption that the problem is psychosomatic. Unlike a group of people with pain, even generalized pain, there's no valid basis to assume that all those people for which there is not yet an explanation are all suffering from the same condition and should be evaluated and treated in the same way. Both the inclusion criteria and treatment strategy are flawed.

What if PACE had come out with a clear positive result with a significant number of patients recovering? ...

Then they would have been able to apparently legitimately claim that CBT/GET works for all the conditions under their broad umbrella Oxford definition. The fact that those with ME/CFS got worse could be swept under the carpet, especially as they didn't do proper recording of harm.

Exactly

 

[Trish]

For pain there can be the same problem. For example, for some back pain, exercise is beneficial, but for some it isn't. For fibro, I hear exercise is supposed to be beneficial, for ME it isn't. So if all pain patients attending a pain clinic are lumped together and put on a GET program, the results will be meaningless.

 

[Grigor]

PEM is heterogenous anyways. There's no need to study fatigue aka Oxford.

 

[Jonathan Edwards]

What if PACE had come out with a clear positive result with a significant number of patients recovering? If, for example, the majority of patients they swept up with their broad brush had depression, deconditioning, fibromyalgia and other forms of fatiguing condition, and a signficant number had improved their activity levels markedly.

Then they would have been able to apparently legitimately claim that CBT/GET works for all the conditions under their broad umbrella Oxford definition. The fact that those with ME/CFS got worse could be swept under the carpet, especially as they didn't do proper recording of harm.

They would have been able to legitimately claim that the treatment worked for at least some - that's all. And of course if they had monitored deterioration properly that would have come to light and the distinction would have been made. If they had done the trial properly and interpreted it properly all would have been well.

Which brings us back to the fact that the error was to do a trial very badly, not to have broad inclusion criteria per se.

 

[Jonathan Edwards]

For pain there can be the same problem. For example, for some back pain, exercise is beneficial, but for some it isn't. For fibro, I hear exercise is supposed to be beneficial, for ME it isn't. So if all pain patients attending a pain clinic are lumped together and put on a GET program, the results will be meaningless.

I am not sure we know that exercise is beneficial either for back pain or fibromyalgia. The best study I have seen in back pain suggests that the best thing is to let the patient decide what to do (a Dutch study).

Wide categories certainly run the risk of producing results that miss discriminations. But narrow categories leave you with no information on everyone else. I you think you know how your treatment would be working then it is sensible to devise criteria that would pick out cases likely to respond but if you have no clear guidance as to how to select people then it is reasonable to go wide.

The BPS people can well be criticised for having an implausible theory but it is not unscientific to have an implausible theory when nobody has much idea what is going on. It is unscientific not to test your theory.

 

[Simbindi]

They would have been able to legitimately claim that the treatment worked for at least some - that's all. And of course if they had monitored deterioration properly that would have come to light and the distinction would have been made. If they had done the trial properly and interpreted it properly all would have been well.

Which brings us back to the fact that the error was to do a trial very badly, not to have broad inclusion criteria per se.

How could that have been ethical with the risk of harm for people with PEM?

 

[Jonathan Edwards]

How could that have been ethical with the risk of harm for people with PEM?

There are always risks of harm in trials. Part of the point of a trial is to make sure there is no harm. Prior to PACE there was no clear evidence of harm as far as I am aware. We still do not have reliable evidence of harm. What we have is a strong suggestion that there may be harm and that is enough to make treatment unethical. The unethical aspect of PACE was to fail very badly to adequately document deterioration.

 

[shak8]

For pain there can be the same problem. For example, for some back pain, exercise is beneficial, but for some it isn't. For fibro, I hear exercise is supposed to be beneficial, for ME it isn't. So if all pain patients attending a pain clinic are lumped together and put on a GET program, the results will be meaningless.

Yes, i keep hearing "exercise" during each visit to my doctor (I have fibro, as well). I want to laugh but the doc writes my pain med prescriptions.

But when half of your body can't do more than one or two 'reps' (no weights at all, no thera bands, nothing) and no eccentric contractions without inducing much unwarranted pain, then the word exercise really doesn't apply. It's careful and very aware movement, careful stretching only after heat and meds, etc. Walking to tolerance, yes, or swimming, yes to increase the endorphins associated with aerbic exercise, but I doubt that many fibro pts get up to that aerobic threshold. Mainly those endorphins come from getting outside or seeing other people in swim pools.


But exercise is touted as palliative in the blurbs out to the GPs in the rather simplistic format as always. Another bias to dismantle. The details are so nuanced.

 

[Simbindi]

There are always risks of harm in trials. Part of the point of a trial is to make sure there is no harm. Prior to PACE there was no clear evidence of harm as far as I am aware. We still do not have reliable evidence of harm. What we have is a strong suggestion that there may be harm and that is enough to make treatment unethical. The unethical aspect of PACE was to fail very badly to adequately document deterioration.

I worked out very quickly that exercise harmed me, back in the mid 90's, so would never have agreed to participate in a trial of GET. I am sure there are many other PWME like me, so any trial would never have been representative of patients. I am unsure how far back the ME charities were doing surveys on this though.