Patients diagnosed with ME/CFS also fit systemic exertion intolerance disease criteria

[user9876]

The one week vs. six months is not duration of symptoms! It is window during which symptoms were reported. So PACE's use of Fukuda required 4 symptoms during the same week, while normal Fukuda allowed 4 symptoms spread out over 6 months. Clearly the PACE definition is a much more serious disease. Here is the quote from the PACE Recover paper:


Nothing was said about the duration of symptoms. I'm sorry you misunderstood the paper in this way, but it is clear that the PACE team used a tightened version of Fukuda, not a loosened version.

But it is a tightening of criteria in order to define recovery hence it results in a loosening of the recovery criteria.

That is if you make it harder to meet the criteria then you get more people who won't meet it which you then define as recovered?

More generally they did something strange with the definition of Oxford in that they added an additional factor that the person had to also meet the trial entry criteria to have passed as Oxford diagnosed. In effect this increased the number of people who counted as not oxford within the data. I think it roughly doubled those who counted as not oxford diagnosis in their recovery figures.

 

[joshualevy]

But it is a tightening of criteria in order to define recovery hence it results in a loosening of the recovery criteria.

That is if you make it harder to meet the criteria then you get more people who won't meet it which you then define as recovered?

Wrong, for two reasons:
First, the discussion was comparing Oxford to other criteria (not a discussion of recovery). I was arguing that they are all the same. PACE found the same results in Oxford and "Fukuda". Since they were really using a harder version of Fukuda, as you point out, that means the fact that PACE found the same results for Oxford and harder-Fukuda is even stronger evidence that all these definitions are almost the same: Oxford, Fukuda, ICC, CCC, SEID, etc. Since harder-Fukuda would be closer to ICC and CCC than regular Fukuda.

Second, although we were not discussing recovery, if we were, then you would be wrong to say that the harder-Fukuda definition means that more would have recovered, because the same definition was used both at the start and end to measure recovery, so tightening (or loosening) the definition would have had little or no impact in overall recovery.

 

[joshualevy]

If you need very large numbers to prove an effect is statistically significant, it is because the effect is inherently weak.

That's one reason, but there are lots of other reasons why you might need more people, and two of them apply very strongly to ME/CFS:

1. Natural fluctuations in the disease: If a disease naturally has good days and bad days, or naturally gets better or worse over time or because of unknown or poorly understood factors, then you need larger numbers of people, to separate out those natural fluctuations from the effects of the treatment.
2. Co-morbid conditions: If a disease is commonly found with other diseases, then it is often hard to gauge the impact of a treatment on the main disease, and you need more people in the clinical trial.

There are other causes as well, but ME/CFS has these two issues, big time. So it should not surprise anyone that larger numbers are required to show effectiveness, and neither of these things has anything to do with the size of the effect being seen.

... that the treatment (both CBT and this type of GET) is mediated only through cognitive factors, by that I mean how we think about our symptoms and report them to our doctor (or other people) or on questionnaires is changed, but not much else.

Careful! The fact that the effectiveness is "mediated only through cognitive factors" as you put it, is evidence (maybe even strong evidence) that the underlying cause is psychological, and I'm sure you did not mean that.

 

[Snow Leopard]

Careful! The fact that the effectiveness is "mediated only through cognitive factors" as you put it, is evidence (maybe even strong evidence) that the underlying cause is psychological, and I'm sure you did not mean that.

Not at all. To be clear, I'm saying the CBT/GET treatments have no physical effects, they are simply distractions, ways of thinking about symptoms differently - leading to people answering differently on questionnaires. (shown by no objective improvements of function - merely improvements on questionnaires)

I don't consider changes found on questionnaires, but not on objective measures of functioning to be a true effect on underlying health.

 

[user9876]

Since they were really using a harder version of Fukuda, as you point out, that means the fact that PACE found the same results for Oxford and harder-Fukuda is even stronger evidence that all these definitions are almost the same: Oxford, Fukuda, ICC, CCC, SEID, etc. Since harder-Fukuda would be closer to ICC and CCC than regular Fukuda.

The definitions are clearly not the same because the numbers of people in the different groups were different. Taking a sample of the same symptoms over a smaller time period is not the equivalent of taking a different set of symptoms where some things are mandatory. If you were to write down the definitions as a series of clauses then you would see that they are not equivalent.

[Edit] Thinking about it a bit more. If you change the definition to have x out of y symptoms in a week rather than a longer time you may get a subset but whilst symptoms fluctuate I'm not sure how much they come and go rather in general I think they fluctuate in severity. Couple that with people wanting to report the symptoms that generally effect them and I suspect it makes little difference - especially for those who have been ill for a long time. The time period is to say that it is chronic. What may happen is some people who are very mild will drop out of the definition. But with something like PACE with entry thresholds these people would not be included. Hence I'm not convinced it had a great effect. However, this is speculation. Jason has done work comparing the different diagnostic criteria and this is probably a good way to understand. If I remember correctly different criteria seemed to represent different levels of severity.

As a general point it is dodgy to claim to use a definition and then use an alternative (non-standard) interpretation instead. This just causes confusion and is likely to cause results to be mis-applied.

Second, although we were not discussing recovery, if we were, then you would be wrong to say that the harder-Fukuda definition means that more would have recovered, because the same definition was used both at the start and end to measure recovery, so tightening (or loosening) the definition would have had little or no impact in overall recovery.

Ok so I hadn't read the full thread. But the point I was making is that they changed the oxford criteria just for the recovery paper.

There is a separate point about what the different definitions mean and how they should be used. To me they all involve a certain amount of guess work as we don't have a clue about disease mechanism and hence we don't know how best to defined a disease (or multiple diseases). However from a research perspective a definition is different from a clinical perspective. Something like Oxford is happy to have co-morbid psychological issues which makes the outcomes of any trial hard to judge, if you filter out say those with co-morbid depression, I think pace gave figures for that, then you loose statistical power.

More generally, the PACE data is somewhat meaningless anyway. Having interventions that try to change how patients view their symptoms and abilities and then judge the success by asking the patients how they view their symptoms and abilities is a bit like saying did I put the right drug in the bottle. So I would be surprised if there were any different results between different diagnostic techniques.

 

[user9876]

Natural fluctuations in the disease: If a disease naturally has good days and bad days, or naturally gets better or worse over time or because of unknown or poorly understood factors, then you need larger numbers of people, to separate out those natural fluctuations from the effects of the treatment

The idea of natural fluctuations is further complicated in the reporting cycle by giving patients flexibility as to when they fill out forms (I think some were sent by post). I suspect this is likely to be done on better days.

I'm not sure that the statistical methods are designed to cope with fluctuating time series data. I suspect if they were they would require much more data or more regularly sampled data and then use average figures. Prior to any trials I think it would be good to get longitudinal data to look at the fluctuations and understand them. Modelling and simulation could then be used to help understand the way fluctuations could effect results.

 

[Snow Leopard]

Natural fluctuations in the disease: If a disease naturally has good days and bad days, or naturally gets better or worse over time or because of unknown or poorly understood factors, then you need larger numbers of people, to separate out those natural fluctuations from the effects of the treatment.

If noise affects the results that much, then again, it suggests the effect size is weak.

 

[AndyPR]

Mistaken conclusions about systemic exercise intolerance disease being comparable to research case definitions of CFS: A rebuttal to Chu et al.
Leonard A. Jason , Madison Sunnquist, Kristen Gleason & Pamela Fox

ABSTRACT
The recent article by Chu et al. contrasted different case definitions that have been used to describe chronic fatigue syndrome (CFS) and myalgic encephalomyelitis (ME). In particular, their study compared the new Institute of Medicine (IOM) criteria for systemic exertion intolerance disease (SEID) with three other ME and CFS case definitions. We appreciate these investigators attempting to use and operationalize the new IOM criteria; however, we disagree with their main conclusion that the percentage of patients selected by the IOM criteria is comparable to the percentage selected by other research case definitions. This conclusion could potentially encourage investigators to use the IOM criteria for research purposes. In this commentary, we discuss our observations of the Chu et al. article with respect to their methodology, illustrating how the conclusions of an investigation can be influenced by the manner in which case definitions are operationalized.

Rest of article behind paywall at http://www.tandfonline.com/doi/abs/10.1080/21641846.2017.1299079?journalCode=rftg20

 

[MeSci]

I couldn't find this (from another list):

Source: Fatigue: Biomedicine, Health & Behavior
Preprint
Date: August 10, 2017
URL:
http://www.tandfonline.com/doi/full/10.1080/21641846.2017.1362750
Ref:
http://www.tandfonline.com/doi/full/10.1080/21641846.2017.1362780
Ref:
http://www.tandfonline.com/doi/full/10.1080/21641846.2017.1299079

Differences of opinion on systemic exercise intolerance disease are
not 'mistakes': A rejoinder to Jason Sunnquist, Gleason and Fox

Lily Chu(*), Ian J. Valencia, Jose G. Montoya
* Corresponding author

Received 20 Jul 2017
Accepted 28 Jul 2017
Published online: 10 Aug 2017

Abstract

Differences of opinion about how case definitions are operationalized
should not be characterized as 'mistaken.' Despite limitations, our
study provides insight into how systemic exertion intolerance disease
(SEID) criteria perform early in the course of myalgic
encephalomyelitis/chronic fatigue syndrome (ME/CFS) compared to other
case definitions. To facilitate early, accurate diagnosis, research
needs to be performed, ideally, prospectively with subjects who have not
been ill for years. In addition, results of past treatment studies may
not be applicable to many ME/CFS patients because criteria used for
research differed from patients' clinical presentations. To avoid this
potential mismatch between research employing the new clinically-focused
SEID criteria should be promoted, rather than discouraged.

Keywords: Systemic exertion intolerance disease (SEID), myalgic
encephalomyelitis (ME), chronic fatigue syndrome (CFS), diagnosis;
Canadian Consensus Criteria (CCC), Myalgic Encephalomyelitis -
International Consensus Criteria (ME-ICC)

--------
(c) 2017 Taylor & Francis
(c) 2017 IACFS/ME

 

[Dolphin]

The original paper is now open access:
http://www.tandfonline.com/doi/full/10.1080/21641846.2017.1299079

This includes the supplemental material.

 

[Dolphin]

A letter was also published on this:

Frank Twisk (2017): Letter to the Editor, Fatigue: Biomedicine, Health & Behavior, DOI: 10.1080/21641846.2017.1357937

 

[Dolphin]

I have now read the forum publications. It is good that there are public discussions on issues. In terms of the subsequent articles, I was more convinced by Chu et al.'s points on this occasion

 

[Dolphin]

The patients were selected using the CFS criteria, and likely the patients were also selfselected (participants of a Stanford study, which takes ME and CFS seriously, thereby attracting relatively more ME (ICC) patients, while ‘severe cases’ were not able to participate). For more definitive answers population studies are essential.

Seems a reasonable point.

 

[Dolphin]

As an aside, I thought it was interesting to see a Venn diagram like this

 

[A.B.]

To facilitate early, accurate diagnosis, research needs to be performed, ideally, prospectively with subjects who have not been ill for years.

This seems like an important point. The health situation of patients probably gets more complicated over time in the subset of patients that don't get better.

The assumption that case definitions that require a lot of symptoms are necessarily more homogeneous is also questionable. More symptoms can mean multiple different problems occurring simultaneously which would produce more noise in a data set. Having a major chronic illness probably makes it easier to develop other conditions.