Discussion in 'Latest ME/CFS Research' started by Sam Carter, Jan 31, 2013.
Forgive me if I am a bit slow on the uptake, but have I got this right? They dismiss the results for the 6 minute walk test on the grounds that they would have improved a lot with suitable encouragement, but the encouragement that patients received in both the CBT and GET groups to take less heed of their symptoms, as opposed to the lack of encouragement in the APT and SMC groups, can be ignored when it comes to responses on subjective questionnaires about their symptoms. Clearly my thought processes are confused: I don't have any awards for contributions to science.
One factor is interesting to me as a mathematician: from baseline to 52 weeks, the standard deviation (which shouldn't be used with such skewed data) for fatigue doubles, for physical function increases by 50%, but for the 6-minute walking test only goes up by 10% to 20%. How much does the standard deviation measure susceptibility to encouragement?
Well Graham I ain't so slow now - can recognise the plug in the sink again without the aid of PACE - this is c.... - some form of encephalitis struck (ie infection in the nervous system). Lets get on with real medicine and science. And never ever apologise in the face of this c... - they have nothing of your modesty and honesty. In short the most twisted self interested bunch one hopes to avoid. (Yep I did psychology at Uni too and can see their twistings - just a game to them). May they all get it - ME.
Summary of Author's reply, with my brief comments
This is intended to supplement my summaries of the letters criticising the trial (which now also includes the original and revised recovery criteria for comparison).
My comments [inside square brackets] mostly draw on the excellent posts by biophile:
The deﬁnition of recovery from any chronic illness is challenging.
Self-rating bias is possible, but inconsistent with higher score for CBT than APT (Pacing), despite higher expectations for APT.
[note they do not attempt to explain the issue raised that the substantial gains in self-rated physical function are not matched by similar gains in the 6MWT]
6MWT was not carried out according to standard practice: patients were not encouraged, and were told they can slow down and stop if needed. Also only a 10m corridor walk was practical, vs the recommended 30-50m. So comparisons with other illnesses are not valid.
[They chose to use 6MWT as an objective test of physical function, it’s a shame they didn’t implement it properly. However, a study of the effect of encouragement on the 6MWT showed it only increased score by about 30m, compared with the 200m PACE trial participants were short of healthy norms at completion. The 10m corridor space is unlikely to make that a big difference. And of course, the conditions were the same for the baseline and final test so the lack of progress (CBT) or minimal progress (GET) still needs explaining.]
It’s not possible for patients to ascribe Recovery to any treatment, could be eg regression to the mean.
As CFS defined by a patient’s reported symptoms, recovery should be measured by multiple reported symptoms rather than a single performance test eg 6MWT
[It’s disappointing the authors lack any interest in comparing their new definition of recovery with either objective measures or patients’ own view of recovery]
Some participants were retired or not working when they fell ill, 6 months post-therapy may be too short a time to affect benefits or employment, therefore benefits/employment data ‘not useful’ .
[Note that Benefits (and insurance) support actually increased during the trial. Also, the study measured hours worked, and some participants were working part-time, so it would be reasonable to see an increase in hours even at 6 months, which wasn’t seen despite the substantial 'recovery' rate].
They agree that the current data only support remission, not recovery - hence they defined Recovery as Remission.
We believe that our approach of using multiple self-report measures provides a reasonable approach to inform clinicians' and patients` choice between available treatments.
[Note that the authors’ letter failed to even address the issue of substantially weakened criteria for fatigue, physical function score and self-rated improvement raised in several letters. It appears their belief in their own weakened criteria runs deep.]
However measured, CBT and graded exercise therapy more likely to lead to recovery, when added to specialist medical care (SMC), compared to either adding APT or SMC alone. 7% recovered in SMC only group according to their recovery criteria – 22% (3x as many) recovered after receiving CBT or GET plus SMC.
[A 22% recovery rate 6 months post therapy would actually be modestly impressive for a debilitating chronic condition such as CFS - if it meant recovery in the sense most patients, or indeed most people, understand recovery. As it is, it’s not clear what the 22% means at all]
The PACE trial has shown that both CBT and GET are moderately effective, safe, cost-effective, and are more likely to lead to Recovery, when added to specialist medical care (SMC), compared to either adding APT or SMC alone. These treatments should now be routinely offered to all those who may benefit from them.
I've been thinking about what I said in the above post, re 'political rehabilitation' and, continuing my recent theme of mild rants, and letting off steam, I was about to post some controversial philosophical thoughts about it, but decided to place them in a blog instead of this thread...
If anyone is interested, I've blogged it here:
CBT for CFS is a political tool, and not a treatment. Discuss.
Or, if I were a cynic, and we all know I'm really a total believer, we could say that PACE was a load of fictional wishes dressed up to look as if it had a real foundation, funded by sources who are pushing their own agenda. Isn't that a pretty accurate description of a political election campaign by those who want power? Of course, I would never think that of people who win awards for science.
I had a quick look into the effect of corridor length on the 6MWD score.
One of the sources cited by White et al states that, "The walking course must be 30 m in length. [...] A shorter corridor requires patients to take more time to reverse directions more often, reducing the 6MWD. Most studies have used a 30-m corridor, but some have used 20- or 50-m corridors (52–55). A recent multicenter study found no significant effect of the length of straight courses ranging from 50 to 164 ft, but patients walked farther on continuous (oval) tracks (mean 92 ft farther) (54)." http://www.thoracic.org/statements/resources/pfet/sixminute.pdf
"Recent study suggests that there is no significant difference on the distance walked in courses between 15 and 50 meters, although oval tracks can increase the results (Weiss et al. 2000)." http://kennisbank.hva.nl/document/219224
These above statements are referring to the following (unpublished?) paper:
Weiss RA, et al. Six minute walk test in severe COPD: reliability and effect of walking course layout and length. Paper presented at ACCP Conference; September 2000; San Francisco.
However I did find another very similar paper (Sciurba et al 2003), possibly the same study, which consisted of 761 COPD patients with severe emphysema: "Among the 14 clinics with straight walking courses, the mean length of the course ranged from 50 to 164 feet (mean 99.9 ± 34.1 feet). [...] Among the 14 clinics with straight course layout, there was no statistically significant effect of track length on 6-minute walk distance in bivariate as well as multivariate analyses."
This small study (Veloso-Guedes et al 2011) of 10 patients on the waiting list for liver transplantation, mentions the American Thoracic Society's recommendations of 30m but found no significant difference between 30m and 20m. Perhaps a larger sample size would find a statistically significant difference, but it probability would not be a large effect? http://www.ncbi.nlm.nih.gov/pubmed/21620120
This small study (Ng et al 2011) on 26 patients with chronic stroke tested the 6MWT with 10m, 20m, 30m. It did find a statistically significant difference, but I have not seen the full text. http://www.ncbi.nlm.nih.gov/pubmed/21530729 Simon later found the full-text and patients walked about 17% further on a 30m track compared to 10m.
The lack of standardisation of the 6MWDT makes you wonder why they didn't think about including actometers in the trial, for a true objective assessment of the participants' activities... Oh, hold on... erm...
I also had another quick look at the learning effect on 6MWD scores.
In a study on 1,514 COPD patients, which performed two 6MWTs on subsequent days, "the vast majority of patients improved significantly in the second test by an average learning effect of 27m".
This commentary on that COPD study explains the significance of learning effects, which can account for roughly 10% (more examples given). http://erj.ersjournals.com/content/38/2/244.full
In the PACE Trial, not only was walking the most commonly chosen activity for the GET group, but participants were encouraged to start at for example 5 minute moderate walks almost daily and gradually increase it. Therapy lasted 24 weeks, and although follow-up was another 28 weeks later, participants were encouraged to maintain the program until reaching 30 minutes of moderate intensity exercise 5 times per week. Regardless of any genuine (but apparently minor if anything) improvements to physical capacity as a result of GET, the learning effect is almost guaranteed for the GET group.
White et al state that some of their trial participants were of retirement age so therefore could have skewed employment figures. However, we were not told how many such participants were in the trial. At the beginning of the trial, the mean(SD) age was 38+/-12 years. Although we do not know the exact distribution of age, it seems highly unlikely that a large proportion were in retirement age. It would not have been difficult to exclude them from analysis anyway.
And according to a UK Government website ( https://www.gov.uk/retirement-age ) : "Default retirement age (formerly 65) has been phased out - most people can now work for as long as they want to."
Yes, and don't forget that they introduced the 6 minute walk along with several questionnaires after the original protocol was accepted, then dropped just the actometer because there were too many assessments.
And just a reminder in case anyone missed it, that they still paid for the actometers, and used them at the start of the trial. It was just at the end of the trial, at the time one measures whether treatments/therapies worked or not, that they dropped using them.
Regarding CBT and GET and return to work, I know it wasn't a target for the Trial, but we have raised the matter and it was also raised in the letters as being a reasonable expectation for 'recovery'.
Afraid I haven't been able to read through this document from the NHS entirely; but you boffin-types might extract from it further detail than I
Occupational Aspects of the Management of Chronic Fatigue Syndrome: a National Guideline
It was published in 2006 by the NHS - with Sharpe and White as external advisors - and on the team were other familiarities. However, it was trying in part to discover by delving through all the published literature at the time; what evidence there was for either therapy - or others - improving work outcomes or I think if other 'treatments' could do better.
I still think it wrong of White to conclude that 'recovery' should not equate with a positive change in employment status/number of hours worked/ and/or education. But alright PACE may not have been set-up to record this as an outcome. However, I think he would say - as this review above concludes I believe - that CBT and GET are the only 'treatments' that offer the best chance of helping for which there is 'credible evidence'.
But that is only if you stick your head in the sand. The whole 'credibility' issue that we spend a lot of time talking about is one area that deserves exposure and these kind of statements that make it into peoples minds rarely say how other management strategies have not been properly trailed; and the conclusions reached about CBT and GET are actually unremarkable and conclusions unrealistic.
For example - one of the studies I believe in the above review concluded that 5 years after 'treatment' a significant proportion of those completing CBT were in employment compared to those who did relaxation: but there had been no continuous 'treatment' throughout this period.
Another example - that people learn for themselves and the illness fluctuates and that natural 'recovery' is possible. It is not credible to say that a therapy or management technique is the cause of an increase in employment status or in the same way and use of language that you might a drug (although not all drugs result in an analysis of the individual to see the biological effect of course outside of Trial).
My own views of CBT and GET and GAM when properly delivered in practice are probably different to others here; but I still maintain that the PACE Trial was overly ambitious to begin with and failed (with some predictability) to live up to expectations.
What we need is for something better - a better Trial or some better treatment - to come along.
I have done my exercise for the day. Off to bed now
This comment on the McCrone et al. (2012) paper:
gives a brief overview of the evidence.
What was the reason again? Must have been a lot of data but they surely knew this going into the Trial. It would still prove valuable to research I would think even now. They should make this available - presuming they kept the things in use for the duration...
The reason they gave was an ethics committee asked them to reduce the work for participants. However, I think they could have reduced some of the large numbers of questionnaires instead.
The decision may have been influenced by emerging data from the Netherlands and the US (Fred Friedberg) (in early 2000s) that CBT/graded activity programmes might not increase total activity as measured by actometers. That is speculation and the authors certainly haven't said that. The authors had used the SF36 Physical Functioning and Chalder Fatigue Questionnaire in previous research so could probably be more confident they would say CBT and GET helped.
They still plan to use the actometer data from baseline in a predictors paper, as far as I know.
- learning effect
- encouragement effect
- priming effect
- barely measurable improvement off a very low baseline...
Not much therapeutic value left over once all these are factored in.
The only 5 year followup study I'm aware of is this one: http://ajp.psychiatryonline.org/article.aspx?articleid=175215 . It has quite a few issues, such as being a chronic fatigue study (oxford definition), not an ME/CFS study. Plus the CBT included pacing, and the "relaxation" was of the sort using muscle tensing - something I've crashed from when guided to do it by a really stupid "relaxation" therapist.
It's also a very small study (53 patients total), and showed improvement in only certain areas: the GET group was working more hours, but had the same amount of unemployed patients as the "control" group (about 50%). In addition, three of the CBT group had been misdiagnosed (1 with cancer, 2 with celiac), which also likely skews the results, as those 3 account for 12% of the CBT group.
Something I find very interesting, if looking at Table 1, is that the CBT patients reported a huge improvement at 6 months on some questionnaires, which completely disappeared at 5 years. Brainwashing, anyone? In fact, the only questionnaire outcome measurement where the was a (barely) statistical signficance at 5 years was self-rating of improvement. And if you take into account the 12% of known misdiagnosed CBT patients (who probably had a full recovery due to getting appropriate treatment), there's probably no statistical significance in any of the 5 year results.
That study is worthless.
I mentioned in this thread that they've avoided updating this at the due date with the evidence from PACE:
Review Date: 2011Key findings of the review:• Cognitive behavioural therapy and graded exercise therapy have been shown to be effective in restoring the ability to work in those who are currently absent from workA large multicentre RCT, PACE is currently under way in the UK, comparing standardised specialist medical care with CBT, GET and pacing. PACE will include work outcomes.
As Dolphin pointed out, the evidence for CBT and GET improving outcomes in employment is amazingly weak, and have now been undermined by the more compelling evidence from PACE.
I don't think that PACE was overly ambitious, but that the psychosocial approaches promoted by White, Chalder, etc need to be able to produce clear and impressive results before they can be justified. Even if they had led to statistically significant improvements in employment rates, or objectively measured levels of activity, the way in which they have been promoted and applied would still be morally unacceptable. Also, I think it should be recognised that the way in which psychosocial interventions for CFS have been promoted has had an affect upon how patients with CFS are viewed, and that this is likely to make it harder for CFS patients to be able to find understanding and sympathetic employers, regardless of whether they received CBT/GET or not. The social harm done as a result of the biopsychosocial approach to CFS is something which cannot be captured in RCTs, unless the control groups take place in different societies.
edit: PS too much interesting stuff in this thread which I've not had time to think about/post on (the title of Bob's blog makes me think that it will correspond with some of the things I've been thinking), just posting now as I'd started a thread on that report in the past.
PPS: Thanks a lot to biophile for looking in to the significance of the hallway length for the 6mwt.
There are so many assertions/implications in a typical paper that it's easy to slipping in to assuming that the trivial ones are true without taking the time to check. It seems like almost all the White ones we do check turn out to be misleading!
Also, regardless of this, the same process occurred for SMC and SMC+CBT, so we can still say that CBT led to no additional improvement in their chosen objective measure of physical fitness.
You can also try a Google Site Search
Separate names with a comma.