Study evaluating NICE, Oxford, and Fukuda prevalence

That one looks like a garbled misquote. It begins to look as if Crawley makes it up as she goes along.

 

To be fair, in the review I thought that they had legitimately extracted what they thought was clinically most relevant for identifying children with ME/CFS. We don't actually know if any particular way of identifying children with a homogeneous syndrome is better than another. I think the review is focused on clinical management so research criteria are not particularly relevant.

 

I think any sort of numbers game is fraught here.

Maybe the main reason for referring to a specialist would be that the GP cannot decide. So in that case diagnosing as something else would be entirely in order and not an indication of an 'error'.

 

Crawley was on the 2007 NICE guideline committee, so perhaps she thinks she owns the description in the NICE guideline and can bend it the way she wants.

It seems that most of the hardcore BPS'ers are using NICE criteria as an alternative to the increasingly discredited Oxford criteria. See for examples Chalder's most recent.

I suspect (and think this was @Medfeb's reason to ask about this) that the 'NICE criteria' are the second broadest case definition and that they are a means to keep focusing on patients with merely chronic (4 months) fatigue when the Oxford criteria are no longer an option.

 

But these are the NICE guidelines for the "diagnosis and management of chronic fatigue syndrome/myalgic encephalomyelitis (or encephalopathy) in adults and children." Therefore, I'd expect the guidelines to clearly tell medical providers how the disease should be diagnosed because they need to make decisions about which patients should be managed by these guidelines. If NICE is not intended to provide diagnostic criteria for clinical care, then how do clinicians decide who has "CFS/ME" and who does not. Not trying to be flippant but it seems so fundamental to diagnosis and management guidelines that guidelines spell out how a provider would diagnose patients.

This seems more than just clumsy wording. It appears NICE has a different definition of PEM than all the other definitions I've seen since NICE considers post-exertional exacerbation of symptoms to be optional and not a part of PEM. That's significant as it could impact both a clinician's ability to diagnose PEM and also their understanding of how it should be managed.

Yea, you are right and I agree BPS researchers appear to be moving on to NICE - it is vague enough for them to select the broad range of patients that have been selected by Oxford in the past. Hopefully, the NICE committee will tighten up those criteria to more specifically and correctly describe the hallmarks of the disease.

 

That assumes you have clinics.

 

I'd say that the majority of referrals that I come across are due to the GPs misguided belief that these centres offer worthwhile treatments.

 

Pretty much everyone who is not a doctor and has not spent three years in clinical training trying to make diagnoses tends to think that way. But one of the things I used to teach the final year students and interns is that if you think about it this is not the way we make decisions in our real lives. 'Making a diagnosis' is really all about probabilities. Is it useful to think that a patient possibly or probably or definitely belongs to a well known pattern of illness with certain implications for prognosis and treatment?

'Diagnostic criteria' are never going to be the right way to do this because they do not take into account the crucial importance of your confidence that any one criterion is even present (often uncertain) and your confidence that if present it is likely to be relevant. 'Diagnostic criteria' is really a hang over from the nineteenth century when naming was all doctors could do. It is essentially a defunct concept. What we now have are classification criteria for research. Criteria are not used in the clinic. What matters in the clinic is covering all relevant possibilities. So a judgment of 30% possibility that an illness will turn out to be lupus is just as relevant to scheduling the next appointment as 100% certainty it is already lupus.

Maybe an analogy would be useful. I am a birdwatcher and I have about a dozen different field guides to European birds. Each gives diagnostic features of redpolls. Some mention mealy redpolls and lesser redpolls. Some mention common redpolls. Most also mention arctic redpolls and a few divide these into Hornemann's and Coue's arctic redpolls. According to the Handbook of birds of the World these are all the same species - redpolls - but there are subspecies and how many is disputed. Many birders still treat arctic redpolls as a separate species.

But crucially, none of these guides gives diagnostic criteria for any type of redpoll or redpolls as a whole.

The reason for that is that a decision about category for any one bird depends critically on the weighting to be attached to each feature. Arctic redpolls in the UK are uncommon visitors from high arctic. A pale redpoll in July is almost certainly not an arctic. One in January could well be if really pale, but in April much less likely. A redpoll with an unstreaked rump could be an arctic but probably not unless the rump is pure white and the rest of the bird is also pale.

The 2017-18 winter we had a bird locally that 100 birders agreed, after photographing and watching for hours, was a Coue's arctic redpoll. That means that we could be pretty sure that the bird had come from a particular sector of the high arctic and would return there. But as to whether arctic redpolls are a separate species or Coue's is 'real subspecies' we knew these questions have no factual answers. These are convenience concepts, just as in medicine. Concepts that allow one to make useful predictions.

I was at the scooping meetings for the NICE committee and there were comments about diagnostic range for the guidelines. These came more or less exclusively from people who have not actually been in the position of being a doctor diagnosing illness. I think they are not just a distraction but a hindrance. For sure we want to identify important features like PEM but we do not want 'criteria'.

I think all this stuff has been raised because of the red herring that maybe GET or CBT work for people who do not have real ME. The reality is that we do not know if they work for anyone, so trying to separate patients into categories is not required for treatment guidelines. The problem with PACE is not that it used Oxford criteria but that it was an uninterpretable trial design. For sure, it makes sense to be extra careful with exercise with people with PEM but everyone gets that.

 

which criteria do you (this community) accept as valid/true

- criteria for mecfs / seid
- criteria for PEM

another question:
has anyone on this forum been diagnosed with cfs me seid according to the NICE criteria ?

perhaps we could have a poll
... with the criteria and the values we got measured, if any
including lung function tests
and heart tests (eeg or what that is)

 

Then why does NICE spend all this effort developing guidelines, including those for people who have "ME/CFS"? How do they decide who should be treated by the ME/CFS guidelines if they don't have some 'features" by which to identify the disease? What is the point of the presentation section if not to provide guidance on diagnosis? And how do they decide what treatments should be included in the guidelines for "ME/CFS" to begin with?

Call them features or required symptoms or whatever you will but certainly doctors must use something to decide whether a patient has a sleep disorder, an autoimmune disease, cancer, ME/CFS, etc so they can decide on the proper course of action. The problem with the Fukuda based guidance here - and as far as I can tell, the NICE guidelines - is that they are little more than a wastebin into which to put medically unexplained fatigue and that does a disservice to everyone.

Regarding why I raised this originally - I was just looking for a link to a study that reported different rates of prevalence by Oxford, Nice, and Fukuda for something I am working on. No red herring.

Regarding PACE - I agree with you that one set of problems is the study conduct issues, conflict of interest, lack of reporting of harms. But the use of non-specific cohort selection methods that don't require core features of the disease is also a problem. It's not an either/or situation. Both are problems. IMO, also calling out the cohort selection problems does not detract from the other issues but rather further amplifies just how screwed up this all is. I appreciate that others see it differently.

Regarding "it makes sense to be extra careful with exercise with people with PEM but everyone gets that" - Unfortunately, lots of people, including in the medical community don't get that. That's one of the reasons that I think its so important to have an accurate definition of PEM and provide guidance to medical providers on how to recognize PEM and the implications of PEM for clinical management plans.

 

Disease experts here recommend pacing (as patients know it) as the foundation of a management program. That form of pacing is about staying inside the energy envelope to try to avoid or at least minimize PEM and the associated crashes and exacerbation of symptoms. Because that's specific to the feature of PEM, the management approach would not necessarily be appropriate for Oxford fatigue or other conditions.

 

As I explained above there are certainly features of the disease, just as there are features of arctic redpolls. But you have to use these features intelligently to come to a decision as to whether or not someone is usefully labelled as having ME/CFS or probable ME/CFS or possible ME/CFS.

The research literature is supposed to provide evidence of treatment efficacy for well defined cohorts. Doctors are then again expected to extrapolate intelligently from these cohorts to individual cases based on what they think are the probabilities of them being representative of that cohort. It is all a matter of intelligent inference of probabilities. Most of the time we do that unconsciously, whether treating patients or identifying birds or buying tomatoes or whatever. The human brain makes useful decisions based oral sorts of contextual information.

What the NICE guidelines gives is a broad list of features that should make the physician consider the diagnosis. If the physician is experienced in the condition they can then use their intelligence to decide. If not the list provides a useful screen for people that would be suitable to refer on to a specialist with a query of a diagnosis. It would be entirely wrong to suggest that doctors use the guidelines alone to make a diagnosis. It would be as irresponsible as suggesting that you start driving a car just by reading the manual. You have to learn the practicalities.

And the guidelines have to cover a broad range because people who have chronic fatigue but not necessarily ME have to be covered by some sort of guideline and the GP has to know roughly where to look for both types of patient. So the guidelines have to have a broad remit. They are a practical guide for doctors in real life.

The best I can think of is the Nacul study but I don't know whether criteria they looked at. I don't think they did 'NICE criteria' because I don't think such a thing actually exists. It would have to be a population study not a clinic study because clinics only take a skewed referral cohort.

I am not sure that it matters. If you set up an uninterpretable study you cannot interpret it for whatever group you recruit.

I agree that if a decent study was set up it might be good to narrow down the cohort, but this is not necessarily true. The reason why not specifying PEM is a specific problem is that it can lead to recruitment bias for studies of exercise. That is a pretty subtle problem. I think it confuses the issues around NICE guideline review. Treatment studies are not necessarily best restricted to specific diagnoses. I have run trial with wider recruitment and for good reason.

The problem with this is that if we say it is important to define PEM so that people with PEM are screened out from GET then it is all too easy for GET to be included in the wider remit that the guideline has to have for other cases of chronic fatigue. We have to go with the realities. If GET is included at all then inevitably it will end up being used the way it is now. From my perspective the only sensible plan is go with the real evidence - which is that there is no reason to recommend GET in any situation. So although exercise intolerance is worth documenting and building into advice to individuals it is not sold as the arbiter of whether people get exercises.

In other words the two key things that need to go into NICE guidelines are strict adherence to evidence and pragmatically. We have to have something that doctors can use. We also need to make sure that a desire for tidiness does not end up having the opposite effect to that hope for.

 

It would have been a problem if the study had actually found that CBT or GET was effective for some patients. But as the study found that nothing worked, that should have been the end of it. The study cannot make any recommendations about CBT or GET (or at least it shouldn't have), so it is irrelevant whether the group of patients was well-defined. There are enough anecdotal reports that GET causes harm to demand that GET cannot just be applied on a "let's try it and see" basis.

But in one sense you could be right. If it turned out that CBT was actually effective for people with a particular flavour of ME, but that these were under-represented under the Oxford criteria because they no longer had fatigue as their main symptom, we would have dismissed a potential treatment (but don't tell the authors that: still, they probably wouldn't follow the logic). Mind you, I don't believe that for one minute.

 

i havent really looked into... but i read in some statement, that the pace trial data were inflated.
but the exercise / the progam clearly helped some patients.

considering the other study with the chronotropic incomptence then indeed for at least a subset the pace/get thing or cbt should be helpful.

there are many user/patient complaints about unhelpful criteria and definitions.
(fukuda/ccc, oxford, nice, perhaps some more)

what are the diagnostics criteria suggested by you, the ill patients ?

wouldnt that be a helpful thing to make such a list ?

here, were sick and educated ppl get together ?

it could be like a list of symptoms, rated as "very strong, "strong", "medium" indicators?

 

Ah, now that statement could cause a vast amount of comment. In brief, CBT and GET persuaded patients to tick boxes in questionnaires saying that they were not as bad as they said originally, but any real test (walking, climbing stairs, return to work) showed no change. As is usual in such studies, some folk did do a bit better, some did a bit worse: that is the nature of the illness, it isn't an indication that CBT helped them.

Suppose I invent a special mind control technique for improving scores when rolling dice. I split a group into two, and teach one of them the technique. They get really enthusiastic, and are convinced it is working. Then the two groups get tested, and, of course, there's no difference between the groups. Can I then just pick out the ones in the trained group who rolled higher scores and argue that it worked for them? Obviously not: that's just random fluctuations.

 

So are you saying that the NICE guidelines are intended to cover all causes of chronic fatigue? What about chronic fatigue due to cancer or heart disease or other conditions? Or is it just for unspecified chronic fatigue? I get that people with unspecified chronic fatigue need to be covered by some guideline. But its not clear why the ME/CFS guideline should cover all cases of unspecified (presumably) chronic fatigue or alternatively why ME/CFS would be stuffed into guideline for unexplained chronic fatigue.

Yea - I know of that study - its a great study. It looked at Fukuda, CCC and one other criteria but not NICE or Oxford so I was looking at comparisons of prevalence about those criteria as well. I agree that those studies lack the quality of the Nacul study and appreciate this is at best a swag

I agree. But conversely, if you do a great study but lack specificity around the cohort selection, then one should question the applicability of the results to a more specifically defined population. And if the study blows it on both counts, that's really messed up. I know we disagree on this but in my opinion, both are an issue. Its not an either/or.

I'm not just focused on whether people who experience PEM should not be prescribed GET although clearly they should not. The issue is broader and goes to the intended "scope" of the condition called ME/CFS in the NICE guidelines and also who gets included as an "ME/CFS" patient in research. What is meant by "ME/CFS." Numerous reports (e.g. the 2006 Gibson Inquiry, the IOM, the 2011 NIH State of Knowledge, the 2015 NIH Pathways to Prevention report, the 2016 AHRQ Addendum) have highlighted the problems in both research and clinical care caused by non-specific definitions.

I guess others may see this differently but I'd hope we are at a point where we can agree that ME/CFS is characterized by a systemic intolerance to exertion (aka PEM) and other symptoms that are required for a diagnosis. Yes, we need biomarkers but in the meantime, these core features could and IMO should be used to both tighten up research cohorts and also focus clinical guidelines on this disease. Maybe this is not what you are saying but it seems crazy to accept a guideline for "ME/CFS" that is so non-specific that it also covers all (or is it just the unspecified) cases of chronic fatigue.