Join the National PR Campaign for ME: Power to the Patient (P2tP)
Have you had enough of all the neglect and abuse of ME/CFS patients? Gabby Klein says now is the time for a National PR Campaign for ME/CFS to impress a change. Join the Patient Revolution to restore power to ME patients ...
Discuss the article on the Forums.

Prevalence and predictors of recovery from chronic fatigue syndrome in a routine clinical practice

Discussion in 'Latest ME/CFS Research' started by Tom Kindlon, Aug 30, 2014.

  1. Tom Kindlon

    Tom Kindlon Senior Member

    Messages:
    310
    Likes:
    1,083
    [The recovery claims continue. Trudie Chalder was one of the principal
    investigators in the PACE Trial where one could have normal physical
    functioning with a SF-36 score of 60 i.e. disimprove on the measure
    and be counted as recovered! This sounds like it may use a similar
    threshold. I would be very interested in seeing a full copy. Tom]

    http://www.sciencedirect.com/science/article/pii/S0005796714001429

     
    Sean, Esther12, barbc56 and 1 other person like this.
  2. Tom Kindlon

    Tom Kindlon Senior Member

    Messages:
    310
    Likes:
    1,083
    This journal doesn't seem to take letters to the editor based on its instructions for authors:
    http://www.elsevier.com/journals/behaviour-research-and-therapy/0005-7967/guide-for-authors

    I also searched PubMed for letters from this journal
    and all it came back with was the following:


     
  3. barbc56

    barbc56 Senior Member

    Messages:
    1,836
    Likes:
    1,381
    I almost choked with laughter the first time I read this statement. Then I realized the word "choice" refers to the medical community and not the patients. Then I felt like crying.

    Thanks @Tom Kindlon

    Barb
     
    Dr Speedy, jimells, Keela Too and 6 others like this.
  4. Esther12

    Esther12 Senior Member

    Messages:
    5,713
    Likes:
    7,357
    Thanks. Definitely something I'd like to see the full paper of.
     
    alex3619 likes this.
  5. Alea Ishikawa

    Alea Ishikawa <>< Senior Member

    Messages:
    196
    Likes:
    274
    "Recovery was defined as no longer meeting Oxford or CDC criteria for CFS measured at 6 months follow-up."

    That might be an issue. The 1991 Oxford criteria is rather broad with "severe, disabling mental and physical fatigue" >50% of the time for 6 months, and includes some other conditions, such as depression, bipolar, and perhaps even standard grieving after a loved one died. Even the 1994 Fukuda criteria is more exclusive.

    I think SOC said it pretty well here, in the last part of her post:
     
    SOC likes this.
  6. Little Bluestem

    Little Bluestem Senescent on the Illinois prairie, USA

    Messages:
    2,803
    Likes:
    2,121
    Midwest, USA
    So a treatment of choice has an approximate success rate of 18% even when using crummy criteria. :wide-eyed::jaw-drop::vomit::bang-head:
     
    Dr Speedy, Min and Revel like this.
  7. alex3619

    alex3619 Senior Member

    Messages:
    8,734
    Likes:
    16,342
    Logan, Queensland, Australia
    Claimed success rate. Claimed. Give therapy, change someone's attitudes, then ask them how they feel about it. Do NOT measure any objective outcomes like exercise capacity, engagement in work or social activity and so on. Do NOT admit that studies that use objective evidence show it is a failure. Do NOT admit that the statistical cutoff for normality (in at least the PACE trial) is based on invalid statistical inference on an invalid statistical data set. Oops, oops and more oops.
     
  8. A.B.

    A.B. Senior Member

    Messages:
    1,065
    Likes:
    3,544
    Of course the success rate would be 100% if patients only wanted to get better and were willing to confess their sins. :rolleyes:
     
    Dr Speedy, Scarecrow, zzz and 8 others like this.
  9. Snow Leopard

    Snow Leopard Senior Member

    Messages:
    2,755
    Likes:
    2,848
    Australia
    No control group and no objective evidence of improvement (eg recovery could just be wishful thinking), means that this study is automatically inconclusive, regardless of the result.
     
    Sean, Keela Too, Min and 3 others like this.
  10. Valentijn

    Valentijn Activity Level: 3

    Messages:
    7,836
    Likes:
    13,188
    Amersfoort, Netherlands
    This entire paper is a hot sticky mess. It makes PACE look reputable by comparison :jaw-drop:

    Criteria for "fully recovered" included a score of 65 or higher of SF-36 physical functioning, and feeling "better" or "much better". In one location it is listed as being "much better" or "very much better" for recovery, but everywhere else it's "better" or "much better". Meaning we can assume some people felt slightly better but generally not "very much better" or "much better", since they don't break down those results. And I'd assume everyone else did "very much worse" since they also withhold that data.

    Even more hilarious is that the p value used for the grand claims in the abstract regarding "predictors" are p = 0.1. I have never seen a study try to pretend that p = 0.1 is statistically significant ... it's absurd.
     
    Last edited: Sep 3, 2014
    Sean, Dr Speedy, Keela Too and 8 others like this.
  11. DanME

    DanME Senior Member

    Messages:
    236
    Likes:
    912
    Hello my friends,

    I ve just read the full study and I am in horror about the statements, the authors make. It is basically the old "it is all in our head and our fault" crap.

    They lump Oxford and Fukuda together and state the two criteria were basically „similar“, but nevertheless had „no complete overlap“ (I see a contradiction here). Their main model is the old hat of avoiding activity after a virus. They say, CBT and GET were proven to be the most effective treatment in CFS, but full recovery was rare.

    They had 200 non-randomised patients, diagnosed by NICE guidelines (having fatigue for the last 4 month) and assigned them to CBT sessions. The CBT was based on „fear avoidance“ and „unhelpful believes“. The patients had to to self report scales including a „perfectionism scale“ and a „pain acceptance scale“ (which was invited for CFS patients, yeah).

    They state, according to Oxford, Fukuda and the self reports, 18% had a full recovery and 60% felt significantly better after 6 month.

    Negative outcome was predicted by disability, catastrophising, depression, fear avoidance, unhealthy believes about emotions, an „all or nothing behaviour“, perfectionism, high parental expectations and embarrassment avoidance.

    Other gems, I found: they state, some CFS patients had not been as healthy as the mean population, before they got sick. So they changed „to be recovered“ to be 1 SD from the population mean (nice bending of statistics here).

    And they cite a study, which stated, only(!) CFS patients with childhood trauma had low cortisol levels (what?).

    So in conclusion, they found exactly, what they predicted in the introduction. The patients were not randomised, they was no control group (maybe with some other treatment idea, like how to accept, that I am severely ill and how make the most of it?). They don’t mention biomedical research and findings, but say, they was no evidence to a pathognomonic manifestations (PEM?). They have no objective criteria at all. They cite quite old studies (mainly from 1990 to 2005). They lump three different criteria together with self reports. I am not a statistician, but I guess, you shouldn’t do that. They bend the statistics, saying most of the CFS patients were not as healthy as the mean population, before they got sick. But have no evidence for this claim. I also doubt the validity of some of the self report scales. And at the end it is all the patients fault, who could be better, if they just believed to be better.

    I am embarrassed by all of this (hey, no embarrassment avoidance). This is not science. It is just a heavy confirmation bias. How can people see severely effected ME patients and still believe in the „activity avoidance“ claim? How can you be a doctor and have a PhD and do such a study? Basic study design? Basic statistics? The central theorem of falsification?

    Alright. What confuses me the most, that this study comes partly out of Haukeland in Bergen, where Fluge and Mella do their fantastic work. How is this possible?
     
    Sean, Dr Speedy, Scarecrow and 6 others like this.
  12. DanME

    DanME Senior Member

    Messages:
    236
    Likes:
    912
    Oh, yeah, I was so angry reading this, I didn't recognise the p value mistake. I cannot believe it.
     
  13. A.B.

    A.B. Senior Member

    Messages:
    1,065
    Likes:
    3,544
    Psychobabble is misanthropic belief, not science. No matter what, it is ALWAYS the patient's fault. The poor therapists wouldn't have anyone to look down upon otherwise.

    The case of Gert Postel shows that being a skilled psychobabbler is enough to make a career in the mental health field, even where there is no substance at all behind the babble. He literally invented diagnoses on the spot, and nobody suspected a thing.
    http://en.wikipedia.org/wiki/Gert_Postel
     
    Last edited: Sep 1, 2014
    alex3619, zzz, Min and 2 others like this.
  14. alex3619

    alex3619 Senior Member

    Messages:
    8,734
    Likes:
    16,342
    Logan, Queensland, Australia
    This dataset is exponential, and the median for working age people is 100. Mean is inappropriate to make inference from, as this includes sick people. Some CFS patients are indeed not as healthy as the general population prior to onset, or at least diagnosis ... so what? The relevance of this has yet to be established.

    Any use of SD on an exponential data set is invalid. No conclusion drawn based upon that has any validity. Did they get advice from any statistician? Did they listen? Did they ask the right questions? Did they bother looking at the data set?

    I think that if its correct they used a p value of 0.1 as cutoff this answers my questions. This paper is statistically nonsense.
     
    zzz, Min and Valentijn like this.
  15. Dolphin

    Dolphin Senior Member

    Messages:
    7,271
    Likes:
    7,897
    The one time I saw p<0.1 being used was by promoters of CBT in the CBT, CFS and Grey Matter Volume study by the Nijmegen team (De Lange et al. 2008) http://brain.oxfordjournals.org/cgi/pmidlookup?view=long&pmid=18587150

    [I have also seen p<0.1 being used in multivariate analysis]
     
    Last edited: Sep 2, 2014
    Esther12, alex3619 and Valentijn like this.
  16. Dolphin

    Dolphin Senior Member

    Messages:
    7,271
    Likes:
    7,897
    Study used the so-called empiric criteria (Reeves et al., 2005) covering 2.54% of the population. Waste-of-time criteria in my mind which bring in all sorts of psych patients.
     
    Valentijn likes this.
  17. biophile

    biophile Places I'd rather be.

    Messages:
    1,485
    Likes:
    6,315
    Loose criteria, non-randomized cohort, no control group, dependence on self-report in a study of perception-altering therapy, and p value cutoff of 0.1. What could possibly go wrong? Chalder, where are the PACE results for predictors, it's been over 4 years already since you've had the data? In addition to problems already highlighted by others, this paper also has several other errors and discrepancies, once again showing that peer review is inadequate. Here are just a few ...

    In some places, the SF-36 cutoff is ≥65, in others it is >65.

    Deale et al did not use 1 SD from the mean to establish thresholds. Unless it was cited for a population mean.

    Knoop et al did use 1 SD from the mean but used a healthy (not general) population as this paper does (big difference).

    This is misleading. Jenkinson et al's paper was about a working age population where 1 SD below the mean score is about 70, and 65 is only within 1 SD of the mean for males and females aged 55-64 years.

    Deale et al do not use Likert scoring for the CFQ either; it used the much stricter bimodal score of <4 as the cut-off point, and this was established arbitrarily in a previous paper using ROC analysis, not mean SD. Deale et al do not mention anything about being within 1 SD of the population mean, but do cite Garratt et al for physical function norms:

    "A cutoff score of 83 was used to denote good outcome, as it represents the ability to carry out moderate activities (e.g., carrying purchases, moving furniture). This cutoff was used in the original trial (3) and is similar to the mean in the U.K. general population (9)."

    9. Garratt A, Ruta D, Abdall M, Buckingham J, Russell I: The SF36 Health Survey Questionnaire: an outcome measure suitable for routine use within the NHS? Br Med J 1993; 306:1440–1444

    The actual population mean in the Garratt et al paper is 79.2. No SD is given but can be crudely estimated as 22.

    The involvement of Chalder (PACE) could explain the mean SD sloppiness. Maybe this is Flo's first paper on CFS? Yet, as bad as this sounds, at least they have attempted to justify the use of mean and SD, usually there is no explanation. I doubt CFS researchers have ever sat down seriously with population data to statistically evaluate what should represent recovery. Somewhere along the way CBT researchers have conveniently latched onto mean SD, and once it gave easy and absurdly low recovery thresholds allowing inflated recovery rates, it has become more common.

    A physical function score of 65 is much higher than where I am now, but it is also insulting to claim it is "fully recovered". Considering my background and age, "fully recovered" is at least 95 and is a major difference to 65. Perhaps incumbent politicians could hire CBT researchers to redefine economic recovery at the next recession or GFC downturn.

    We deserve better than this. Would CBT researchers be happy taking a pay cut or salary equal to 1 SD below the mean income of a general population that included the unemployed and retired?

    So much for the alleged spiraling descent into increasing disability, then?
     
    Last edited: Sep 4, 2014
  18. Sean

    Sean Senior Member

    Messages:
    1,654
    Likes:
    4,194
    We deserve better than this. Would CBT researchers be happy taking a pay cut equal to 1 SD from the mean income of a general population that included the unemployed and retired?

    He he he…

    This is the sort of pointed analogy/metaphor that we need to use.

    The assumptions underlying the BSP model really don't stand up to even modest scrutiny, do they?
     
  19. alex3619

    alex3619 Senior Member

    Messages:
    8,734
    Likes:
    16,342
    Logan, Queensland, Australia
    Fully recovered for most is 95+. Only with serious comorbidities could you claim otherwise.

    I think many economists already run with the same playbook. We keep being told things are fixed. Yet I am waiting for the next and bigger economic crash. The last one was only a warmup. Nothing is fixed. Governments are afraid to take the difficult measures, and instead focus on scapegoating groups like the disabled.

    We need lots of little anecdotes, aphorisms and quips. Then use them a lot.
     
    jimells likes this.
  20. Snow Leopard

    Snow Leopard Senior Member

    Messages:
    2,755
    Likes:
    2,848
    Australia
    The median of a healthy population in their 30s is about 95, I'd say >85 for a young population and >80 for a older population is a fair cut off for 'back to normal'/'recovered' on the SF-36 PF scale.
     
    Valentijn likes this.

See more popular forum discussions.

Share This Page