• Welcome to Phoenix Rising!

    Created in 2008, Phoenix Rising is the largest and oldest forum dedicated to furthering the understanding of and finding treatments for complex chronic illnesses such as chronic fatigue syndrome (ME/CFS), fibromyalgia (FM), long COVID, postural orthostatic tachycardia syndrome (POTS), mast cell activation syndrome (MCAS), and allied diseases.

    To become a member, simply click the Register button at the top right.

Rehabilitative therapies for chronic fatigue syndrome: a secondary mediation analysis of PACE trial

Bob

Senior Member
Messages
16,455
Location
England (south coast)
I've just come across this on Twitter, via Firestormm.
See the highlighted text...

Is the chronic fatigue syndrome an exercise phobia? A case control study.

Gallagher AM, Coldrick AR, Hedge B, Weir WR, White PD.
2005
J Psychosom Res. 58:367-73.
http://www.ncbi.nlm.nih.gov/pubmed/15992572

Abstract

OBJECTIVE:

The aim of this study was to test whether patients with chronic fatigue syndrome (CFS) have an exercise phobia, by measuring anxiety-related physiological and psychological reactions to ordinary activity and exercise.
METHODS:
Patients and healthy but sedentary controls were assessed over 8 h of an ordinary day, and before, during and after an incremental exercise test on a motorised treadmill. To avoid confounding effects, those with a comorbid psychiatric disorder were excluded. Heart rate, galvanic skin resistance (GSR) and the amount of activity undertaken were measured, along with state and trait measures of anxiety.
RESULTS:
Patients with CFS were more fatigued and sleep disturbed than were the controls and noted greater effort during the exercise test. No statistically significant differences were found in either heart rate or GSR both during a normal day and before, during and after the exercise test. Patients with CFS were more symptomatically anxious at all times, but this did not increase with exercise.
CONCLUSION:
The data suggest that CFS patients without a comorbid psychiatric disorder do not have an exercise phobia.
 
Last edited:

Dolphin

Senior Member
Messages
17,567
From the abstract of the PACE Trial protocol: http://www.biomedcentral.com/1471-2377/7/6

Mediators and moderators will be explored using multiple linear and logistic regression techniques with interactive terms, with the sample split into two to allow validation of the initial models.

There was no sign of the sample being split in two.

It's not 100% clear from the text whether the random split was just for predictors (baseline variables) or also process variables.
Predictions and process of treatment
Associations between post-treatment outcomes and both predictor and process variables (including demographic, illness duration, and other putative clinical indicators) will be examined using multiple linear and logistic regression modelling techniques, including a limited examination of interactions both amongst pairs of predictors and between predictors and treatment groups. We anticipate that the sample size will be sufficient to identify important general predictors from a random-split, training set of two thirds (~400), with partial validation in the remainder, used as a test set. Shrinkage techniques (to allow for over-optimism in variable selection) will be applied in the development of a prognostic model to be applied to participants outside the trial.
 
Last edited:

Dolphin

Senior Member
Messages
17,567
From the PACE Trial protocol:

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

Process variables
1. Step test of fitness 43]
Used

2. Borg Scale of perceived physical exertion 44]
Used

3. The symptom interpretation questionnaire 34]
Used

4. Exercise and activity scale
Not reported in Chalder et al. (2015)

Please indicate how much you agree with the following statements by ticking the
appropriate box. (Please tick one only per line)

Strongly disagree
Disagree
Neither agree nor disagree
Agree
Strongly agree

Belief

I should avoid exercise when tired

Doing less helps fatigue

Exercise is harmful

I should avoid physical activity



5. PHQ symptom sub-scale

Not reported in Chalder et al. (2015)

During the past 4 weeks, how much have you been bothered by any of the following
problems?

Year Month Day Fore. Midd.Sur. Centre Participant
Date completed

Not bothered at all
Bothered a little
Bothered a lot

a Stomach pain
b Back pain
c Pain in your arms, legs, or joints: knees, hips, etc.)
d Menstrual cramps or other problems with your periods [Women only]
e Headaches
f Chest pain
g Dizziness
h Fainting spells
i Feeling your heart pound or race
j Shortness of breath
k Pain or problems during sexual intercourse
I Constipation, loose bowels, or diarrhoea
m I Nausea, gas, or indigestion
n Feeling tired or having low energy
0 Trouble sleeping

6. HADS scale combined score
Not reported (they reported HADS Depression and HADS Anxiety separately)
 

user9876

Senior Member
Messages
4,556
If anyone has more info on the Cognitive Behavioural Responses Questionnaire (CBRQ), it'd be appreciated.

I've just been looking and I'm struggling to find much so not a very popular scale. I was searching for CBRQ and fear avoidance to try to find details of the sub-scales they quote. The only references given in the lancet paper (the appendix) seem to be

5. Knudsen AK, Henderson M, Harvey SB, Chalder T. Long-term sickness absence among patients with
chronic fatigue syndrome. Br J Psychiatry. 2011;199 (5):430-1. Epub 2011/09/10.
6. Skerrett TN, Moss-Morris R. Fatigue and social impairment in multiple sclerosis: the role of patients' cognitive and behavioral responses to their symptoms. J Psychosom Res. 2006;61(5):587-93. Epub 2006/11/07

The first reference (5) see comments below just refers to the second and hence isn't very useful. I've not managed to find an available copy of the second paper. From what I've seen I'm assuming its not a well used scale and the sub-scales are certainly not well used or tested. I was wondering about any claims of linearity since this is mentioned as an underlying assumption for the statistical methods used in the PACE mediators paper. I was also concerned that there may be overlap between the questions used in the sub-scales which may have an effect if there are assumptions of independence for the mediator variables.

I've still to look in the PACE protocol and will try to do that tonight.


Long-term sickness absence among patients with chronic fatigue syndrome
http://bjp.rcpsych.org/content/early/2011/08/26/bjp.bp.110.082974.full.pdf
The Cognitive and Behavioural Responses Questionnaire (CBRQ) is a new scale designed to assess patients’ cognitive and
behavioural responses to symptoms. It has been validated on 230 patients with CFS (further details available from T.C. on
request) and has been used in patients with multiple sclerosis. Previous factor analysis of Likert-scored data revealed five cognitive
subscales and two behavioural subscales. Of the cognitive subscales, one measures the level of symptom focusing, and four assess how
patients interpret their symptoms (catastrophising, damage beliefs, fear avoidance and embarrassment avoidance). The two
behavioural response subscales measure all-or-nothing behaviour, and avoidance/resting behaviour.

In the paper she says more is available in the appendix but it just says:

Description of the cognitive and behavioural subscales assessed by the Cognitive and Behavioural Responses Questionnaire (CBRQ)
- Fear avoidance: avoidance of activities due to fear of worsening symptoms (e.g. I am afraid that I will make my symptoms worse if
I exercise)
. Catastrophising: catastrophic cognitions regarding chronicity of symptoms (e.g. I will never feel right again)
. Damage beliefs: beliefs that symptoms and symptom severity reflects true damage to the body (e.g. the severity of my symptoms must mean there is something serious going on in my body)
. Embarrassment avoidance: avoidance of social situations due to feelings of shame and embarrassment over symptoms (e.g.
the embarrassing nature of my symptoms prevents me from doing things)
. Symptom focusing: cognitive preoccupations on symptoms (e.g. when I am experiencing symptoms it is difficult for me to think of anything else)
. All-or-nothing behaviour: behaviour characterized by periods of high activity resulting in overextension and subsequent prolonged
periods of resting (e.g. I tend to do a lot on a good day and rest on a bad day)
. Avoidance resting behaviour: excessive resting and avoidance of activity (e.g. I stay in bed to control my symptoms

The reference given is to a paper by
 

Simon

Senior Member
Messages
3,789
Location
Monmouth, UK
Is this really mediation?
I already queried whether the authors have actually shown mediation, because
- the putative mediators and outcomes change at the same time, so it isn't possible to say which affects which.
- the putative mediators 'mediate' in SMC and APT groups too, even though those therapies don't target such beliefs (a few in APT, none in SMC)

However, in the case of the 6MWT, described as "a potent mediator of GET", the position is clearer: the protocol clearly described it as an outcome, not a mediator:
Secondary outcome measures – Secondary efficacy measures
...
8. The six-minute walking test will give an objective outcome measure of physical capacity [31].
And simply from a rationale prespective, it seems likely that for GET the mediator is the agreed programme of increasing exercise (walking was most popular), and walking distance in six minutes is the logical outcome, not the mediator.

So of the two mediators highlighted by the authors in the abstract, one appears to be an outcome and the other 'avoidance behaviour' changes in tandem with the outcome of fatigue/function so mediation can't be proven.

Curiously, an earlier mediation study from the PACE team concluded
How do treatments for chronic fatigue syndrome work? Exploration of instrumental variable methods for mediation analysis in PACE
There was modest mediation of CBT and GET effects (approximately 20% of the total effect).
Which contrasts with the 'up to 60%' mediation effect seen in this new study. It would be interesting to know what's changed.

Just for fun, this is what the protocol had to say about their expectations of what would predict outcomes.
Predictors of outcome
Predictors of a negative response to treatment found in...
So far, none of those has shown up....
 
Last edited:

Dolphin

Senior Member
Messages
17,567
Simon said:
Just for fun, this is what the protocol had to say about their expectations of what would predict outcomes.
Predictors of outcome
Predictors of a negative response to treatment found in
previous studies include having a mood disorder, membership
of a self-help group, being in receipt of a disability
pension, focusing on physical symptoms, and pervasive
inactivity [3,18,19]. There is however no general agreement
on which are the most important predictive factors.
I think there will probably be a separate predictors paper (i.e. looking at baseline characteristics and whether one can tell from them who will or won't respond.).
 

anciendaze

Senior Member
Messages
1,841
Any chance one of us might get a paper published with the title, "A Media-Nation Analysis of the Results of the PACE Trial"? This would naturally focus entirely on the way the trial was reported in various U.K. media outlets in comparison to actual published scientific data.
 

Dolphin

Senior Member
Messages
17,567
If anyone was so inclined, they could post a response on the BMJ website. It posts the vast majority of responses (although it's much harder to get into the print edition).

http://www.bmj.com/content/350/bmj.h227


Research News

Tackling fears about exercise is important for ME treatment, analysis indicates

BMJ 2015; 350 doi: http://dx.doi.org/10.1136/bmj.h227 (Published 14 January 2015)

Cite this as: BMJ 2015;350:h227

Ingrid Torjesen

Author affiliations


Reducing worry that exercise or activity will worsen symptoms of chronic fatigue syndrome (myalgic encephalomyelitis or ME) is important for the success of cognitive behavioural therapy or graded exercise therapy in reducing fatigue and improving physical function, concludes a new analysis of data from the PACE trial, reported in the Lancet Psychiatry.1


Earlier results from the PACE (adaptive pacing, graded activity, and cognitive behaviour therapy—a randomised evaluation) trial, published in the Lancet in 2011, showed that in a group of 641 patients cognitive behavioural therapy (CBT) and graded exercise therapy (GET) were of more benefit at 52 weeks to people with chronic fatigue syndrome than adaptive pacing therapy (APT) and usual specialist medical care.2


In GET the patient undertakes a personalised and gradually increasing exercise programme delivered by a physiotherapist, whereas in APT the patient adapts their activity levels to the amount of energy they have.


Trudie Chalder, professor of cognitive behavioural psychotherapy at King’s College London, said that although the original PACE trial results indicated that CBT and GET offered safe and effective treatment options for most patients with chronic fatigue syndrome, the improvements seen were moderate. “By identifying the mechanisms whereby some patients benefit from treatment, we hope that this will allow treatments to be developed, improved, or optimised,” she said.


The researchers used mediation analysis to identify the factors—such as physical fitness (as measured by a walking test) and beliefs about activity—through which CBT and GET achieved their beneficial effects on fatigue and overall physical function. These potential mediating factors were measured at 12 weeks, halfway through the patients’ 24 weeks of treatment, except for the walking test, which was assessed at 24 weeks.


Of the factors they analysed, the researchers found that a reduction in fear that exercise or activity would make the symptoms worse (fear avoidance beliefs) had the largest mediated effect on fatigue and physical function in the case both of CBT and GET (standardised effect ×10 for fatigue, CBT versus APT: −1·22 (95% confidence interval −0·52 to −1·97); GET versus APT: −1·86 (−0·80 to −2·89); for physical function, CBT versus APT: 1·54 (0·86 to 2·31); GET versus APT: 2·35 (1·35 to 3·39)). These factors accounted for 51% of the overall effect on physical function for GET and 37% for CBT, when compared with APT. The respective proportions were 61% and 34% for the same comparisons for the effect on fatigue.


Increase in exercise tolerance (metres walked during a six minute walk) was a potent mediator of the effect of GET, when compared with APT (standardised effect ×10 for fatigue: −1·37 (−0·76 to −2·21); for physical function: 1·90 (1·10 to 2·91)), but not CBT.


Other factors, including avoidance of activity, damage beliefs, and catastrophising, were weaker mediators of the effects on fatigue and physical function.


Chalder said, “Our results suggest that fearful beliefs can be changed by directly challenging such beliefs (as in CBT) or by simple behaviour change with a graded approach to the avoided activity (as in GET). Clinically, the results suggest that therapists delivering CBT could encourage more physical activity such as walking, which might enhance the effect of CBT and could be more acceptable to patients.”


The original 2011 study attracted much criticism over its definitions of secondary outcomes, protocol changes, and generalisability, and the results were disputed. Charles Shepherd, an honorary medical adviser to the ME Association, said that evidence from patients collected by the society indicated that “most people with ME/CFS find CBT to be ineffective (or even harmful); around half consistently report that GET makes them feel worse; and the vast majority report that pacing is the most acceptable and effective form of activity management.”3


Chalder said it was important, because individual experiences of the condition were highly variable, that exercise was personalised to the patient and then introduced gradually. “The most important thing is that right at the very beginning when you do the assessment you find out what the person is able to do and then work from there,” she told a press conference.


“On a day when they feel less symptomatic and very energetic they may do more than they are probably able to do and then of course suffer as a consequence. What we try and do is level out their activities so that [they are] more even and more distributed across the week and the day.”


Notes

Cite this as: BMJ 2015;350:h227


References

↵Chalder T, Goldsmith KA, White PD, Sharpe M, Pickles AR. Rehabilitative therapies for chronic fatigue syndrome: a secondary mediation analysis of the PACE trial. Lancet Psychiatry 14 Jan 2015, doi:10.1016/S2215-0366(14)00069-8.

↵White PD, Goldsmith KA, Johnson AL, Potts L, Walwyn R, DeCesare JC, et al, for the PACE trial management group. Comparison of adaptive pacing therapy, cognitive behaviour therapy, graded exercise therapy, and specialist medical care for chronic fatigue syndrome (PACE): a randomised trial. Lancet2011;377:823-36.CrossRefMedlineWeb of Science

↵ME Association. Response to PACE trial recovery paper. 15 Feb 2013. www.meassociation.org.uk/2013/02/me-association-response-to-pace-trial-recovery-paper-15-february-2013.
 

Scarecrow

Revolting Peasant
Messages
1,904
Location
Scotland
Chalder said it was important, because individual experiences of the condition were highly variable, that exercise was personalised to the patient and then introduced gradually. “The most important thing is that right at the very beginning when you do the assessment you find out what the person is able to do and then work from there,” she told a press conference.

“On a day when they feel less symptomatic and very energetic they may do more than they are probably able to do and then of course suffer as a consequence. What we try and do is level out their activities so that [they are] more even and more distributed across the week and the day.”
That sounds an awful lot like pacing to me. Did someone just reinvent the wheel?
 

Bob

Senior Member
Messages
16,455
Location
England (south coast)
BMJ said:
Chalder said [...] “On a day when they feel less symptomatic and very energetic they may do more than they are probably able to do and then of course suffer as a consequence. What we try and do is level out their activities so that [they are] more even and more distributed across the week and the day.”
So, by Chalder's own admission, patients are not afraid of activity on their less symptomatic days, whereby they are "energetic" and "do more than they are probably able". Chalder seems to be trashing her own paper and theories.
 
Last edited:

Dolphin

Senior Member
Messages
17,567
That sounds an awful lot like pacing to me. Did someone just reinvent the wheel?
UK graded activity-oriented CBT involves a period of stabilisation and then, supposedly, some increases.

At both stages, patients are encouraged to take regular rests.

If CBT helps people a bit, I have wondered whether it may be more to do with the regular rests than with the graded activity element of it which may never happen (the PACE Trial results don't suggest increased activity in the CBT group).
 

Bob

Senior Member
Messages
16,455
Location
England (south coast)
I've just been looking at the therapists' manual for GET, and it's interesting that there are pages of explanation about the deconditioning theory in it, which is now irrelevant, seeing as they now say that fitness was not a mediator for GET, and fitness levels did not improve after GET.

But what I was looking for was this:
PACE manual for therapists (GET) said:
CFS/ME setbacks usually involve an exacerbation of their symptoms, leading to a significantly reduced functional capacity. Participants may describe these as a ‘relapse’, or ‘crash’. People with CFS/ME can usually identify an increase in physical activity which may have attributed towards their setback. Sometimes setbacks also appear to be caused by sleep disturbance, a new active infection or emotional distress. It is normal, and likely, that participants will suffer setbacks throughout the GET programme.

If the plan has been undertaken carefully, with a low baseline and small increments as
planned, it is unlikely to be the exercise programme that is responsible. However,
it is important to ascertain whether any components of the GET programme may have
contributed towards setbacks, and to adapt the plan immediately to avoid difficulties
.

So, GET did involve an element of pacing, because therapists were instructed to "adapt the plan immediately to avoid difficulties." i.e. to adaptively reduce activity levels after relapses. Sounds like pacing to me!
 
Last edited:

user9876

Senior Member
Messages
4,556
I've just been looking at the therapists' manual for GET, and it's interesting that there are pages of explanation about the deconditioning theory in it, which is now irrelevant, seeing as they now say that fitness was not a mediator for GET, and fitness levels did not improve after GET.

But what I was looking for was this:


So, GET did involve an element of pacing, because therapists were instructed to "adapt the plan immediately to avoid difficulties." i.e. to adaptively reduce activity levels after relapses. Sounds like pacing to me!

It would be interesting to know how much of the gain occurred in the first few weeks for some people as they find a base line activity level.
Although I tend to think the gain in subjective scores with CBT and GET is from telling people to view their symptoms differently. We don't know what the distribution of score changes was and so it could be that some people who weren't pacing benefited a lot from a base line process.
 

Sidereal

Senior Member
Messages
4,856
This description of GET sounds a lot less insane and nothing like what early GET programmes were like in the 1990s. Those programmes were regimented and were truly based on the belief (of psychiatrists) that symptom exacerbations didn't matter since they did not reflect underlying damage so the activity was supposed to be increased on a schedule regardless of symptoms. Needless to say that kind of GET had excellent efficacy at turning mild patients into housebound/bedridden ones. Nice move to completely change the nature of the therapy and still call it the same thing.
 

chipmunk1

Senior Member
Messages
765
Fear avoidance: avoidance of activities due to fear of worsening symptoms (e.g. I am afraid that I will make my symptoms worse if
I exercise)

that is not irrational it is reasonable.

. Catastrophising: catastrophic cognitions regarding chronicity of symptoms (e.g. I will never feel right again)

not unreasonable considering the poor prognosis

.
Damage beliefs: beliefs that symptoms and symptom severity reflects true damage to the body (e.g. the severity of my symptoms must mean there is something serious going on in my body)

very reasonable assumption
. Embarrassment avoidance: avoidance of social situations due to feelings of shame and embarrassment over symptoms (e.g. the embarrassing nature of my symptoms prevents me from doing things)

the embarrassment is caused by giving patients a psychosomatic diagnosis that is not accepted by society !!

.
Symptom focusing: cognitive preoccupations on symptoms (e.g. when I am experiencing symptoms it is difficult for me to think of anything else)

again a normal human reaction.

.
All-or-nothing behaviour: behaviour characterized by periods of high activity resulting in overextension and subsequent prolongedperiods of resting (e.g. I tend to do a lot on a good day and rest on a bad day)

contradicts point 1. you can't have both.

Avoidance resting behaviour: excessive resting and avoidance of activity (e.g. I stay in bed to control my symptoms

not unreasonable.
 
Last edited:
Messages
73
thanks so much for the analysis on here

Im still trying to get my head around it...

If anyone could correct me if Im wrong but superficially it seems like:

The 6MWT whilst showing statistically significant improvement - the real world improvement for improving the illness is practically negligible
The rest of the evidence for GET and CBT is take from simplistic subjective assessments - and again - is very slight.

Is there a log anywhere for the claims and press released around the PACE announcements? I am interested in making a comparison chart with what they said, what the press printed - and what the real results were...

EDIT

and in case anyone else is confused:

Conceptually, moderators identify on whom and under what circumstances treatments have different effects. Mediators identify why and how treatments have effects.

http://www.psych.ox.ac.uk/publications/139599
 
Last edited:

user9876

Senior Member
Messages
4,556
thanks so much for the analysis on here

Im still trying to get my head around it...

Conceptually, moderators identify on whom and under what circumstances treatments have different effects. Mediators identify why and how treatments have effects.

In trying to understand the techniques they are using to look at mediators I think it is much too strong to say that they have identified how a treatment works. With their model they basically assert that there are a set of potential modifiers in this case due to timing of questionnaires and then they look for correlations between treatment type, the 'mediator' variable and the end point variable. Given a correlation they are saying there is a mediation but without careful checks on time of events then its not really mediation. For example, if the same argument (say telling people to ignore symptoms and they will get better (CBT and GET) encourages people to change answers to activity fear questions at 12 weeks and fatigue or physical function at 52 weeks (perhaps as well as 12/24 weeks but I don't think they checked that) then that comes up in their mediation score as significant even though it is just correlation from a common underlying variable.

They seem to have used a very complex description of the techniques they used which suggests to me they may not get them and have not really thought out the problem in detail. Stats help us think about data it never provides an algorithm where you plug in data and get results. Thought about what is going on and hence what the algorithms say is necessary.
 

Bob

Senior Member
Messages
16,455
Location
England (south coast)
I've yet to read the methodology for the statistical analysis (not really looking forward to it) but i wonder if they carried out the analysis on individual outcomes or on the average outcomes. I assume it must be for individual outcomes if the analysis is to be meaningful?
 
Last edited: