1. Patients launch $1.27 million crowdfunding campaign for ME/CFS gut microbiome study.
    Check out the website, Facebook and Twitter. Join in donate and spread the word!
The Pathway to Prevention (P2P) for ME/CFS: A Dangerous Process
Gabby Klein gives an overview of the P2P process, shedding light on the pitfalls with advice as to what we can do in protest ...
Discuss the article on the Forums.

Pain inhibition and post-exertional malaise in ME/CFS

Discussion in 'Latest ME/CFS Research' started by Dolphin, Apr 1, 2010.

  1. Dolphin

    Dolphin Senior Member

    Messages:
    6,826
    Likes:
    5,941
    Another Belgian exercise study

    http://www3.interscience.wiley.com/journal/123308811/abstract?CRETRY=1&SRETRY=0

     
  2. shrewsbury

    shrewsbury member

    Messages:
    1,540
    Likes:
    55
    So - in my naive way - I see that there are good guys and bad guys doing research in Belgium. And these seem to be the good guys - objective science.

    Really like that they found PEM both with with both types of exercise, and via self-assessment AND through measurement it seems, but I'm not clear what that measurement is. But objective measurement of PEM seems important.

    I always want to know how long after the exercise assessments (subjective/objective) are made.
     
  3. Dolphin

    Dolphin Senior Member

    Messages:
    6,826
    Likes:
    5,941
    Some information from the paper (minus discussion)

    Definition of self-paced (not exciting but so people now what happened):
    Also for period between experiment 1 and 2:
    They say this in the discussion on the finding:
    One way of saying this is that patients don’t “boom and bust” as some psychiatrists, etc like to say.

    *This means what the people doing the test predicted they should do (75% of estimate if good day, 50% if bad day)

     
  4. Dolphin

    Dolphin Senior Member

    Messages:
    6,826
    Likes:
    5,941
    A couple of comments

    They point out in the abstract that
    However, they actually checked for lots of such correlations. It is unclear to me how many. It could be the 19 symptom of the CFS Symptom List x2 (both tests). So the normal statistical test is whether p<0.05/n where n is the number of symptoms. Which would be p<0.05/38=0.0013 which is not reached. [Aside: I also think such testing is very strict but at the same time they should have at least mentioned they didn't adjust. If they did do 38 comparisons then this could easily be a chance finding].

    This is all they say in the Methods - Statistical analysis section about the number of comparisons:

    This is what they say in discussion
    They may have only checked the symptoms where there was a worsening which would mean less than 38 comparisons.

    The reason you do this is can be seen by tossing a coin 6 times.
    The probability of getting all heads with an unbiased coin is (.5)^6=0.015625.
    The probability of getting all tails with an unbiased coin is (.5)^6=0.015625.
    The probability of getting all heads or all tails with an unbiased coin is 0.015625*2=0.03125.
    If one sets a threshold of p<0.05, then one can say this is unlikely to happen by chance.
    However, if one had one hundred people toss a coin each six times, there is a good chance, that one or more will have all heads (or all tails). But one can't surmise that that individual coin is biased - it could be a chance finding because one did so many comparisons.

    ==========
    Point #2:

    This didn't work out as intended for Nijs et al.

    They thought that the patients would have a better response to the "Self-paced and physiologically limited exercise" and so this might be useful in an exercise program.

    However, there were actually more examples of significant worsening after the second test than after the first test!

    The results in each category didn't reach significance:
    but it still is interesting.

    You couldn't recommend a management regime that involved the amount of exercise in the "self-paced and physiologically limited exercise", based on these results.
     
  5. shrewsbury

    shrewsbury member

    Messages:
    1,540
    Likes:
    55
    Oh - I'm getting so confused - so Nijs has taken part in some of the previous "bad" research, but here had to publish the results of "good" research???:Retro redface:

    I so appreciate your last 2 posts Tom - the breaking up of the text alone helps me be able to try to decode the info, and your translations into layman's terms are incredibly helpful for me. And then, the piece de resistance - explaining the relevance of it all in Point #1 - priceless.
     
  6. Dolphin

    Dolphin Senior Member

    Messages:
    6,826
    Likes:
    5,941
    Thanks IF. :Retro smile:

    Nijs is too into exercise for my liking. Apart from physical therapists/physiotherapists who have the illness themselves, I'm afraid I tend to distrust them and have reason to distrust him from several papers in the past.

    That doesn't mean that when they do objective science it can't be good. But they and he can still spin the results a bit - not as much as others but it can still cause problems. Whether he/they knew they should adjust for multiple testing, we'll never know but one sees it enough in studies that if you read enough research papers, you should think of it even if you have forgotten some of your statistics education.
     
  7. Esther12

    Esther12 Senior Member

    Messages:
    5,299
    Likes:
    5,615
    This is relatively interesting for a paper like this - which is a shocking reminder of just how boring these papers are.

    It does seem like there is some really slow and painful progress being made in building up evidence about CFS. Maybe we'll start to get a positive cycle as evidence increases, and study design becomes less driven by the prejudices of the academics involved?

    Thanks for posting the details TomK.
     
  8. jspotila

    jspotila Senior Member

    Messages:
    1,074
    Likes:
    609
    Tomk, I haven't been able to find the full text of this paper. Do you have it in pdf format (or can you point me to a source)? Thanks!!!
     

See more popular forum discussions.

Share This Page