Somebody seeks info on objective results from Graded Exercise Therapy studies

[Pendergast]

Thank you @anciendaze & @Esther12 !
Yes, I suffer from PEM at 24/48h of doing exercise so that is clear to me.

Also, I was aware of the big tons of rubbish in the PACE trial and the great job of Mr.Tuller, so I think I got what you are trying to tell me.

Sadly, due to my cognitive and language limitations it's hard for me to understand well long articles ( that is why I sometimes get lost).

I'll try to read again your messages so maybe I will understand them better. Here it's 4.40AM and I'm like a zombie,but I got your points I think.

One thing I still see problematic is that in drug trials you can blind them even if the outcomes are subjetives too. But in these "we want you to exercise and change your mind"-like studies, the only possibility is to ensure good objetive outcomes and if even that is not well done, then it is very bother

The paradoxical thing is that the PACE trial, due to its flaws, finally showed the opposite they wanted to show, that is: GET & CBT don't work!! But there are still Drs. who think they do. That's very frustrating but I send them the New York Times article, so maybe they will learn....

Thank you all and excuse me for my English and my questions. Let's see if I can sleep something.

A great hug!!

 

[Snow Leopard]

I think all the confusion about CBT and GET would clear up if you assume the whole purpose of therapy is to train patients to complain less about their condition, and train doctors to ignore those complaints.

Cynical, but, well, more truth than they'd like to admit.

 

[Snow Leopard]

But, how can researchers blind these kind of studies? And what kind of objetive measures would be the best if we don't have biomarkers nor clear parameters to test for -yet-?

You can't blind the studies, but that's not what we're suggesting. We're stating that unblinded trials with subjective outcomes are highly prone to bias. This is not a controversial statement - this is why double blinding is the only way to get drug approval without strong and relevant objective outcomes (eg blinding is not required for certain cancer treatments where death is highly probably).

One interesting way of (partially) controlling for the bias is to compare to a drug (which has a mode of action which is highly plausible to patients), along with a placebo arm. The blinded placebo arm measures the placebo effect which can be compared to the unblinded non-pharamacological treatment. Such a study has already been done for CBT for CFS and guess what? CBT showed no effect compared to the placebo group.

 

[A.B.]

One interesting way of (partially) controlling for the bias is to compare to a drug (which has a mode of action which is highly plausible to patients), along with a placebo arm. The blinded placebo arm measures the placebo effect which can be compared to the unblinded non-pharamacological treatment. Such a study has already been done for CBT for CFS and guess what? CBT showed no effect compared to the placebo group.

Can you provide a link?

 

[Snow Leopard]

"Immunologic and psychologic therapy for patients with chronic fatigue syndrome: A double-blind, placebo-controlled trial" 1993
http://www.sciencedirect.com/science/article/pii/000293439390183P

Conclusion: In this study, patients with CFS did not demonstrate a specific response to immunologic and/or psychologic therapy. The improvement recorded in the group as a whole may reflect both nonspecific treatment effects and a propensity to remission in the natural history of this disorder.

And the Chalder/Deale/Wessely reply:
http://www.amjmed.com/article/S0002-9343(99)80332-5/pdf

 

[Pendergast]

We're stating that unblinded trials with subjective outcomes are highly prone to bias. This is not a controversial statemen

Of course it isn't. I totally agree. The bias is not only in the outcomes of such a poor methods but in the researcher's minds before starting the trial too.

And the PACE trial did harm us all over the world. I was told to exercise by some Drs. based on it.

Really shamefull stuff

 

[anciendaze]

Just want to highlight something Esther12 posted about alleged improvements in the "6-minute walk test". The improvements in group mean value would still have left patient's performance grouped with performance of patients diagnosed with stage II heart failure. If you were dealing with such heart patients failure to move to a different stage would mean this would not be counted as a clinically-significant change because there was no change in diagnosis.

This leaves us with a question about statistical significance, and there we discover that 1/3 of the group did not provide data because they declined to perform the test either before or after therapy. We have data from a separate "step test" reported in a separate publication, and this showed no significant change at all. This is a discordant result the authors seem to have buried while promoting their favored therapies.

The simplest hypothesis in interpreting data from the "walk test" is that there was no statistically or clinically significant improvement at all, but that allowing patients who felt worse after therapy because of fatigue to decline the "walk test" biased results by counting patients who felt somewhat better, but omitting some who felt worse.

Their own data strongly suggests a null result, a zero.

I believe we have seen two patients here on the forum who were actually in the GET arm of the study, and believe they suffered set-backs from that trial. How could this be if it was not reported as such? The authors seem to have decided that the adverse events in that group were not the result of their efforts. This allows them to claim that the increase in "adverse events" caused no change in "adverse responses", a new term. This seems to imply those who reported the excess adverse events simply had bad luck, a novel methodological invention from these authors.

 

[joshualevy]

I'm still working on a list of all the studies that have ever compared GET/CBT to other therapies, but this is the list so far. I'm sorry the formatting is so bad:

S1 Crawley 2013 GET / CBT Registry Analysis 800
S2 GETSET [soon] [unknown] Intervention 200
S3 PACE 2011 GET / CBT Intervention [F1][F3] 600+ /4 Oxford
S4 FINE None [F2]
S5 FITNET 2012 CBT Intervention [F4] 120
S6 Núñez 2011 Pacing(*) Intervention 120
S9 Sandler 2016 GET / CBT Intervention 250+
S10 Deale 2001 CBT Intervention 60
S12 Wiborg 2015 CBT Intervention 200
S13 Worm-Smeitink 2016 CBT Intervention 450
S14 O'Dowd H 2006 CBT Intervention 153
S15 Janse 2017 CBT Intervention 195
S16 Schröder 2017 $ STreSS Intervention 120 FSS
S17 Qure 2017 CBT Intervention 150 /3 QFS
Sxx FITNET-NHS [2022] [unknown] Intervention 660

I haven't gone through to see which of these trials used objective measurements. The ones that I know included at least one subjective measures were S3, S14, and S16. The ones that I know were all subjective were S9, S17, S12, and S13. I haven't looked up the rest.

But consider S15, it uses the "Sickness Impact Profile" which is specifically described as "A 68 question objective measure designed to assess quality of life and level of dysfunction that results from disability or illness." but I assume you would not consider it "objective" because it is questionnaire based.

If you look at all (questionnaire based and non-questionnaire based) studies there are 13 studies, the breakdown is as follows:
1 study found no differences between CBT/GET and the other treatment (S4)
1 study found Pacing had fewer side effects, but otherwise no differences (S6), and
11 studies found CBT/GET were better than the comparison group (all the rest).