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!
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.

Adverse events and deterioration reported by participants in the PACE trial of therapies for CFS

Discussion in 'Latest ME/CFS Research' started by Tom Kindlon, Apr 26, 2014.

  1. Tom Kindlon

    Tom Kindlon Senior Member

    Messages:
    254
    Likes:
    786
    Free full text: http://www.sciencedirect.com/science/article/pii/S0022399914001883
     
    Last edited: Apr 27, 2014
    Valentijn and Esther12 like this.
  2. Tom Kindlon

    Tom Kindlon Senior Member

    Messages:
    254
    Likes:
    786
    If anyone is considering a letter, here are the instructions:

    It is hard to keep a 1000 word letter tight so it might be better to aim for lower than the maximum.
     
    Valentijn likes this.
  3. A.B.

    A.B. Senior Member

    Messages:
    721
    Likes:
    2,116
    So why do these results seem to have no relationship to reality? How is deterioration of physical functioning defined?
     
    Min likes this.
  4. Dolphin

    Dolphin Senior Member

    Messages:
    6,840
    Likes:
    5,960
    Possibly circumstantial evidence of response bias in CBT (or some who have done CBT) i.e. they don't report problems.
     
    WillowJ and biophile like this.
  5. Bob

    Bob

    Messages:
    8,591
    Likes:
    11,513
    South of England
    So they've published it at last. The abstract is meaningless. We need the full paper.
     
    alex3619 and Roy S like this.
  6. Esther12

    Esther12 Senior Member

    Messages:
    5,302
    Likes:
    5,624
    I thought that this paper was also meant to be looking more generally at mediators for outcomes... when are we going to get those results - that's what I'd been waiting for! Ah well.
     
  7. Countrygirl

    Countrygirl Senior Member

    Messages:
    855
    Likes:
    696
    UK

    The full paper is there........or am I hallucinating????:)

     
    Valentijn and Esther12 like this.
  8. Bob

    Bob

    Messages:
    8,591
    Likes:
    11,513
    South of England
    :) lol, no you're not hallucinating. It wasn't available earlier.

    It's suddenly become open access!

    A nice bit of bedtime reading!
     
    Last edited: Apr 26, 2014
  9. Countrygirl

    Countrygirl Senior Member

    Messages:
    855
    Likes:
    696
    UK
    I've scanned it, Bob.................from under the safety of the duvet. Now I have the thoughts of the infernal trio bedding down with me tonight and it will take a whole zopiclone and more to beat them down..

    This irrelevant twaddle really is enough to make a man eat his young. Just how much longer do we have to put up with this disgraceful situation? At this rate, I shall have fallen off my perch long before we receive that heartfelt :rofl: apology from the UK medical profession for their half-century of neglect and abuse of the ME community. It is my ambition to live long enough to savour that moment. Will I make it?:cry:

    C.G.
     
    peggy-sue, Min and Bob like this.
  10. minkeygirl

    minkeygirl Narcissism = lack of self awareness

    Messages:
    2,293
    Likes:
    1,483
    Left Coast
    I think that it was put out by the Journal of Psychosomatic Research says it all.
     
  11. Dolphin

    Dolphin Senior Member

    Messages:
    6,840
    Likes:
    5,960
    To answer your second question, a deterioration of physical functioning required a decrease of 8 or more points on a questionnaire, the SF-36 physical functioning subscale (many CBT and GET investigators/authors tend to talk about "physical functioning" without making it clear they're talking about a questionnaire).

    This questionnaire involves 10 questions about activity where people are asked to say one of the following regarding certain activities: "Yes, limited a lot", "Yes, limited a little" or "no, not limited at all". There is a difference of 10 points between "Yes, limited a lot" and "no, not limited at all" and 5 points between the two smaller changes. So the deterioration required was two small steps or one larger steps.

    It seems possible that somebody who became more cautious about not overdoing it, as encouraged by APT, might report this sort of level of change, without it necessarily representing a true deterioration.
     
    Last edited: Apr 26, 2014
  12. Bob

    Bob

    Messages:
    8,591
    Likes:
    11,513
    South of England
    Table 5 seems the most relevant, in terms of how CBT/GET affected patients:
    http://www.sciencedirect.com/science/article/pii/S0022399914001883#t0025

    They looked at deterioration, at one year, as an equivalent measure to improvement (as reported in the main Lancet paper) for self-reported fatigue and physical function, and other measures.

    Roughly 10% reported deterioration in physical function after CBT+SMC (9%) and GET+SMC (11%).
    But this needs to be compared with the SMC-alone group for which the deterioration rate was 18% for physical function.
    So there was roughly an 8% lower deterioration rate in the CBT/GET groups than in the SMC-alone control group. So CBT/GET compare favourably with the control in this paper.

    Self-reported fatigue has fairly similar difference between groups.

    There were 2 or 3 "serious adverse reactions" in each group, out of roughly 160 participants in each group. I can't see what the threshold for a serious reaction was.
     
    Last edited: Apr 26, 2014
    Valentijn and Esther12 like this.
  13. biophile

    biophile Places I'd rather be.

    Messages:
    1,389
    Likes:
    4,665
    I think there are a number of reasons for the discrepancy.

    Clinical trials are generally much better organized and controlled than routine clinical practice. One of the main aims of the PACE Trial was to test the safety of these therapies. The principal authors were already proponents of CBT/GET and convinced these were safe (based on clinical experience and small trials which showed no major worsening in group average scores), so one could say they were motivated to prove that their position and role in the ongoing controversy about safety was justified by evidence. So extra care was put into harms prevention and safety monitoring.

    CBT and GET groups were encouraged to carefully increase activity and tolerate a degree of symptom exacerbations for a week or so as long as the activity levels were sustainable and did not lead to a relapse or reduction in function. This safety net is an important factor when considering that the objective evidence suggests that CBT and GET does not lead to clinically significant improvements in total activity levels, total service use, employment hours, walking distances, welfare dependency etc. There was encouragement to test the boundaries, but no evidence that the boundaries were actually expanded in a sustainable manner, something which CBT/GET proponents tend to spin doctor or gloss over.

    The 'safety' of CBT/GET seems to depend on not actually achieving what proponents claim to achieve with it. CBT/GET outside a RCT may be less respectful with patients and this is what proponents and an AfME survey blame for the adverse effects universally reported in patient surveys. I also think Dolphin's point about response/reporting bias is important if patients were told that they focus too much on symptoms and that they exaggerate the seriousness of symptoms etc.

    Other explanations include the recruitment process. 80% of candidates who were provisionally or definitely diagnosed with CFS before referral were excluded from the trial. The most common reason for exclusion was not meeting Oxford criteria for CFS. This criteria allowed multiple symptoms but demands that fatigue to be the only principal symptom. Some candidates refused to participant or were excluded because they had a preference for one of the therapies. I think it is a safe bet that some candidates did not want to do CBT/GET because they thought it was useless or harmful. The medical assessment may have also excluded ME/CFS patients with organic dysfunction.
     
    Last edited: Apr 29, 2014
    peggy-sue, Cheshire, Roy S and 5 others like this.
  14. Esther12

    Esther12 Senior Member

    Messages:
    5,302
    Likes:
    5,624
    The oversight and safety provisions built into PACE were good. If similar provisions were in place for these treatments in routine practice, and if patients were provided access to unspun data on efficacy (like Pace's protocol defined outcomes, with information on problems with response bias in unblinded trials and the info we have from more objective measures of outcome), then I wouldn't have a problem with them.

    I thought that GET could have more fluctuations in reported symptoms, but overall, I'm not surprised by these results (although i've not read the paper properly, so could well be missing something important).
     
  15. biophile

    biophile Places I'd rather be.

    Messages:
    1,389
    Likes:
    4,665
    No objective evidence that CBT and GET are actually rehabilitative.
    APT had a 70% rule which is not part of general pacing.

    It may not be obvious in the sentence quoted above, but in the Lancet paper there were about twice as many SAEs in the GET group and the difference was statistically significant. It was only after the assessors were unblinded to group allocation that these were judged unrelated to treatment, which seems awfully arbitrary. Maybe it should be the participants judging whether their symptom exacerbations were related to GET?

    Readers may be unaware that the full trial protocol also graded NSAEs by severity, which is not reported in this paper. These could be graded Mild / Moderate / Severe. There has been no attempt to explore the difference.

    http://www.meactionuk.org.uk/FULL-Protocol-SEARCHABLE-version.pdf

    This is important because a NSAE may range somewhere between say, a minor annoyance, to, being bedridden for several weeks. "Transient exacerbation of fatigue or pain, expected as a normal reaction to CBT or GET in patients with CFS/ME, which does not have significant impact upon function (see 14.1.1 (a))"

    Their definition of a non-significant impact upon function is essentially anything that is not a serious adverse event (SAE), which is described in the paper and quoted below, so I will not repeat again here. It is strict and patients may not agree that anything less than a SAE is not a significant impact upon function.

    Unless I misread something, the specific reference to part (a) of 14.1.1 rather than the whole subsection must be a typo because it is "death", which is an absurd threshold for a significant impact upon function!

    They do not seem to mention that they changed the definition of serious deterioration since the protocol was published? i.e. reduction of SF-36 score by 20 or more points from one assessment to two consecutive assessments.

    Once again, overemphasis on fatigue, which is not required of any other CFS or ME definition. Even the CDC criteria does not require fatigue to be the principal symptom.

    In other words, definition of SAE was strict, and there was room for interpretation when judging whether it was a reaction to treatment (therapy or medication). @Bob . A SAR is just a SAE that was judged to be related to treatment.

    No such consideration for non-normal distribution of physical function scores in the population data when questionably and controversially changing the thresholds for "normal" and "recovered".

    So, smart enough to realize that they were probably conned on the issue of "recovery".

    Again, they failed to look at the grading of NSAEs (mild / moderate / severe).

    Maybe some centres / therapists were better at convincing participants that symptoms aren't as serious or severe as they thought? Again, Dolphin's point about response bias is important here. Nevertheless, variations of ascertainment between centres is also an plausible contributing factor.

    Or it may reflect illness characteristics which are incompatible with CBT/GET.

    Cue CPET for future studies? They could have included 6MWD deterioration in this paper.

    Especially when the assessors are unblinded to allocation before arbitrarily deciding whether SAEs are SARs.
    There is no mention of non-serious adverse reactions, which is also mentioned in the trial protocol!

    "Non-serious adverse events or reactions will be assessed by the RN at each follow-up assessment interview. A risk assessment has been undertaken, and we have concluded that the therapies are of low risk to participants. Non-serious adverse events will be reported according to the usual regulatory requirements."

    http://www.biomedcentral.com/1471-2377/7/6

    Can't say I'm surprised, given APT's questionable 70% rule.
     
    Last edited: Apr 26, 2014
    Dolphin, peggy-sue, Esther12 and 5 others like this.
  16. Dolphin

    Dolphin Senior Member

    Messages:
    6,840
    Likes:
    5,960
    I came across one participant from the GET arm of the trial on another forum.

    She said the physiotherapist (=physical therapist) said to her that she could use housework to replace exercise on some days if she wasn’t feeling so good, or something along those lines.

    This is not part of the philosophy of GET which is about adding exercise on top of daily activities.

    Also, housework can be of varying intensities including very low intensity. This is not how GET is supposed to be: the exercise session is supposed to be of the same intensity and either the same length or a bit longer.

    If such measures to “play it safe” were more common across the trial, one could see how fewer adverse events would be reported than with a more rigorous program including one that stuck rigidly to the manuals.
     
    Last edited: Apr 26, 2014
    peggy-sue, Esther12, Roy S and 4 others like this.
  17. Valentijn

    Valentijn Activity Level: 3

    Messages:
    6,520
    Likes:
    9,666
    Amersfoort, Netherlands
    So they only looked for adverse events at three points, with a HUGE gap of 7 months between the 2nd and 3rd assessment. Without being asked at other appointments, and with CBT and GET involving a healthy dose of brainwashing, that seems pretty scanty.
    While they are blinded in determining if an adverse event is serious, it's not clear who is making this determination. If it's anyone involved in the research, or friendly CBT/GET practitioners, there's a general interest in showing no serious adverse events in any arm of the trial. And then they are unblinded to determine if the event is a reaction to the treatment. So basically they get to determine if their pet therapy caused a severe adverse event. Can anyone say "Bias"?
    This ties into the statement at the beginning of the paper, that women and neurotics report more adverse events than manly sane types. Basically, if you report adverse effects, they're implying it's because you're more psychosomatic, female, fat, shy, neurotic, and/or depressed.

    Also they've made a clear error in labeling obese patients: a BMI of greater than 30 represents mere obesity, not "morbid obesity". Morbid obesity requires a BMI of at least 35 AND obesity-caused health problems. If health problems aren't a result of the obesity (there's no indication here that they are), then a BMI of at least 40 is required for the label of "morbidly obese".
    Basically different clinics reported vastly different amounts of non-serious adverse events per patient. If I'm reading the table right, Kings, Oxford, and Bart's reported them at about half the rate as Edinburgh, Royal Free, and Bristol. Aren't Kings, Oxford, and Bart's the more hardcore centers, with heavy involvement by the PACE authors?
    Ah, they finally come out and say it in the Discussion: having too many symptoms means you're psychosomatic. Nice theory, but do they have any proof? Nope, not a single shred of evidence supporting it.
    This part is complete trash, resulting from their insistence in using questionnaires which are grossly inappropriate for patients with physical disability. Apparently not being capable of doing things we used to do, even if we want to do them, means we're depressed. Whoops! Or it just means they're a bunch of idiots. I favor the "idiot" theory - it's much better supported by the available data.
    Indeed, the size is the only strength of the study - but that pretty much disappears when you study fatigue patients instead of ME patients, in addition to all of the other problems. I guess 3 NSAE assessments are an improvement over 0, but it's still pretty damned pathetic.
    Well, at least they admit that the lack of objective outcomes is a weakness. Too bad they never consider it to be big enough of a weakness to rectify it with objective measurements. And frankly, if they did, they'd just take the Nijmegen model and explain that lack of objective improvement simply proves that objective improvement is irrelevant, since the subjective measurements say they must be cured.
    This is the bit where they completely fail to discuss the Serious Adverse Events which they decided, when unblinded, are not reactions to the therapy. They talk about Non-serious Adverse Events at great lengths, give lip service to Serious Adverse Reactions, and do not mention how GET had twice as many Serious Adverse Events as the "controls". Naturally they do not explain how that difference disappeared when the researchers were unblinding to the treatment and realized it was their money-making therapy involved. Oops.
     
    Simon, Dolphin, peggy-sue and 6 others like this.
  18. Valentijn

    Valentijn Activity Level: 3

    Messages:
    6,520
    Likes:
    9,666
    Amersfoort, Netherlands
    A bit more about "Serious Adverse Events" versus "Serious Adverse Reactions":
    I think the 2nd sentence was deliberately written to imply that there was no significant difference between SAEs or SARs per treatment. But I'm doubting that is accurate regarding the SAEs.

    From the initial PACE trial, there were 17 SAEs among 13 of the 161 GET patients (some of the 13 had multiple SAEs). This compares to 7 SAEs among 7 of the 160 SMC patients (each of the 7 had 1 SAE). They don't examine that data too closely, but I can't imagine that having 243% of the SAE rate (2.4 times as many) wouldn't be statistically significant, especially in such a big trial.

    In the Deterioration paper, they completely omit the total number of SAEs per group, but only report the number of patients per group with SAEs. They show the number of participants who had SAEs (13 GET, 7 SMC), and say it isn't statistically significant, with a p-value of 0.15. This seems a little fishy even with those numbers ... any stasticians who can comment?

    They are also being quite dishonest in mentioning "Serious Adverse Events" in the paper, but then displaying the number of participants with such events, rather than the number of events. Even the table which shows the data for participants with SAEs (table 5) does not specify that they're showing the number of participants with SAEs, rather than the number of SAEs. In fact, the label is simply "Serious Adverse Events", which I would think is major error in their paper.

    I'm also rather disappointed in the lack of detail regarding SAEs and SARs. Non-serious Adverse Events are broken down by category, but there is 0 data about the serious events and reactions.
     
    Last edited: Apr 27, 2014
    Dolphin and peggy-sue like this.
  19. alex3619

    alex3619 Senior Member

    Messages:
    7,499
    Likes:
    11,993
    Logan, Queensland, Australia
    Crashes with ME are often temporary, though they don't have to be. For some of us that is a few days to a week or two. This will downplay the impact and in most cases they probably will fail to detect it if measured at two time points.

    Again the PACE papers fail to come to grips with the exercise physiology. I would like to know how their patients fair now using a 2 day CPET, and its a shame we were not aware of CPET much earlier. I think CPET results will probably blow their research into the deep cold of space. It would be nice to see.
     
    peggy-sue, A.B. and Valentijn like this.
  20. biophile

    biophile Places I'd rather be.

    Messages:
    1,389
    Likes:
    4,665
    @Valentijn. Yes, 7 months is a long time between assessments. I struggle to remember how many non-serious adverse events have occurred in the last 7 months or much detail about them, just that some have occurred.

    Re "independent assessors of the trial safety data", from the Lancet paper:
    Hiroko Akagi, Alastair Miller, and Gavin Spickett.

    Good point about the questionable definition of morbid obesity used.

    "Basically, if you report adverse effects, they're implying it's because you're more psychosomatic, female, fat, shy, neurotic, and/or depressed [...] having too many symptoms means you're psychosomatic [...] [diagnosis of depression] resulting from their insistence in using questionnaires which are grossly inappropriate for patients with physical disability".

    Indeed. It doesn't seem to occur to them at all that having more symptoms could alternatively reflect a more severe illness. All the stuff about somatization and depression is a tad circular. As you said, "not being capable of doing things we used to do, even if we want to do them, means we're depressed" (or at least counts towards it).

    Re SAE vs SARs again. I haven't looked at the numbers in closer detail, but I remember last year when they had to comment on the PACE Trial during the NHS online clinic week last year when the experts got too many questions. They tried to claim that were no significant differences in "any" of the safety measures across the four trial arms (except for the measure of satisfaction in the SMC-alone group), which presumably also includes "serious adverse events" (SAEs) since they listed it in the same paragraph among the other safety measures.

    However, this contradicts their 2011 Lancet paper which states that "There were more serious adverse events in the GET group than there were in the SMC group (p=0·0433)." An effect size for this statistically significant difference is not given, but when looking at Table 4, it appears that SAEs were roughly 2.4 times more common (as you said) for +GET, a difference which just happens to disappear when the assessors were unblinded.

    I think someone mentioned this potentially relevant systematic review paper on another thread:
    http://www.bmj.com/content/344/bmj.e1119

     
    Last edited: Apr 27, 2014
    Simon, Dolphin, Sean and 2 others like this.

See more popular forum discussions.

Share This Page