GETSET (white) in Lancet 22/06/17

[Barry53]

Am I right in thinking that GES is different to GET, in that GES is performed in the absence of a clinician, albeit having been previously primed by one? And does the guidance differ from that given for GET?

And they really exhibit a paranoid fear of objective outcomes don't they. So simple yet so hard to include somehow. Can't imagine why

 

[Barry53]

The two primary outcomes were fatigue—measured by the Chalder Fatigue Questionnaire19 (CFQ; Likert scoring 0, 1, 2, and 3; range 0–33; highest score is most fatigue)—and physical function, measured by SF-36 PF (range 0–100; highest score is best function) 12 weeks after randomisation. These questionnaires are valid and reliable measures, and have been used in previous trials

[my bold]

Valid and reliable measures of what though? Certainly not objective indications of fatigue or physical function. What the hell is the point/chance of "measuring" physical function subjectively? You are not measuring physical function, you are measuring a perception of physical function, probably relative to an earlier perception of physical function; not at all the same as real physical function. As with all things White et al, it is all in the mind!

 

[Deepwater]

Something I noticed about the Chalder Fatigue Scale is that, although the range of possible total scores is 0 to 33, the score of someone who reports that all of their symptoms are "no more than usual" compared to when they were last well is actually 11, not 0.

.

Only recently ill people need apply, then? Goodness, after 22 years I can't remember what well felt like - I have this cloudy memory of dancing and walking a long way, and not feeling it. I remember the emotion - the joy of doing it - but not the physical feeling. Can they be sure, in any case, that some of the patients did not mistakenly understand "no more than usual" as meaning what it says rather than "no more than a well person"?

 

[Forbin]

Only recently ill people need apply, then? Goodness, after 22 years I can't remember what well felt like - I have this cloudy memory of dancing and walking a long way, and not feeling it. I remember the emotion - the joy of doing it - but not the physical feeling. Can they be sure, in any case, that some of the patients did not mistakenly understand "no more than usual" as meaning what it says rather than "no more than a well person"?

It is confusing. The instructions imply that they think the questionnaire will mainly be used by the recently ill.

"We would like to know more about any problems you have had with feeling tired, weak or lacking in energy in the last month. Please answer all the questions by ticking the answer which applies to you most closely. If you have been feeling tired for a long while, then compare yourself to how you felt when you were last well. Please tick only one box per line." [Bolding Mine]

http://www.goodmedicine.org.uk/files/assessment, chalder fatigue scale.pdf

 

[Forbin]

Valid and reliable measures of what though? Certainly not objective indications of fatigue or physical function. What the hell is the point/chance of "measuring" physical function subjectively? You are not measuring physical function, you are measuring a perception of physical function, probably relative to an earlier perception of physical function; not at all the same as real physical function. As with all things White et al, it is all in the mind!

It's also important to keep in mind that the 4.2 drop between subjects and controls is not a percentage. The paper never refers to it as a percentage of anything. It would be 12.7% of 33, but they never say that, either. It's just a drop in the mean score between subjects and controls. The precise meaning of that oddly precise difference in scores seems... unclear.

I seem to remember another case where the answer was something like 4.2 - and yet it was unclear what the question was... Hmmm.... what could that have been...?

 

[Sean]

Meh, what's a few T-cells decimal places between friends?

 

[Snow Leopard]

These questionnaires are valid and reliable measures

This is not actually, true - the quality and patient acceptibility have never been reliably assessed. In fact all of the self report questionnaires don't meet best practise on this basis.

https://www.ncbi.nlm.nih.gov/pubmed/21590511

 

[Forbin]

On the quantification of subjective data....

 

[trishrhymes]

On the quantification of subjective data....

 

[trishrhymes]

I have wondered about this with all these trials; is there any record of how long the participants had been ill for (or how long since they were diagnosed)? Or are they recruiting/diagnosing 'on the hoof'?

If I remember rightly, the criteria for entry required at least 4 months fatigue along with a few other symptoms (NICE criteria). I think they said a proportion on the study did not meet the Oxford criteria at the start. Since the only requirement for this is 6 months fatigue, I concluded when I read it that those people had only been fatigued for between 4 and 6 months. I'm quite surprised that in that short time they had managed to get a referral to secondary fatigue/CFS services.

I begin to wonder whether they were having trouble recruiting enough people to take part, so asked GP's to send along other fatigued patients to make up the numbers. The fact that some were 'too high' physical functioning suggests this to me too.

If they were also rejecting anyone who had experienced GET or even read about it and anyone with physical problems with exercising, and accepting people with normal levels of physical functioning, they were clearly not seriously studying ME at all.

And yet, as with PACE despite making the entry criteria as unlike ME as they could manage, they still got such piss poor results. I conclude the same as I did for PACE. This GET nonsense doesn't help anyone - not depressed people, not people with short term post viral fatigue, not stressed people, not people with bad sleep habits, nobody!

 

[slysaint]

I concluded when I read it that those people had only been fatigued for between 4 and 6 months.

Yes, I got the entry requirement, but there doesn't appear to be any record of if any of the participants had been diagnosed before the trials, or how long before.

I'm quite surprised that in that short time they had managed to get a referral to secondary fatigue/CFS services

Yes, quite. Not to mention do all the testing they supposedly are required to do to rule out any other condition, and presumably treat them if they do find anything and see if the 'fatigue' persists.

So it seems they were/are finding 'potential participants' and 'diagnosing' as they went/go along;
no room for bias then

ETA: this question has been answered further down this thread by K Kirke. The figures are on one of the tables in the report.

 

[Large Donner]

These questionnaires are valid and reliable measures, and have been used in previous trials

Science at its best!

"These panels are valid and reliable and have been used on previous towers".

 

[dangermouse]

The "de conditioning" theory and related biopsychosocial (ahem) "treatment" is wrong on so many levels.

 

[Karen Kirke]

Yes, I got the entry requirement, but there doesn't appear to be any record of if any of the participants had been diagnosed before the trials, or how long before.

Yes, quite. Not to mention do all the testing they supposedly are required to do to rule out any other condition, and presumably treat them if they do find anything and see if the 'fatigue' persists.

So it seems they were/are finding 'potential participants' and 'diagnosing' as they went/go along;
no room for bias then

Duration of illness at baseline in the GETSET trial is reported as 46 months for the GES group (that's a median rather than a mean), with an interquartile range of 23-114 months (aka most fell within that range) and 42 months for the control group, with an interquartile range of 25-99 months. This is in table 1, p.6 of the paper. So GETSET participants were generally sick for 3-4 years at the start of the trial, but some for shorter and some for longer.

For comparison, duration of illness in PACE was reported as 32 months with an interquartile range of 16-68 months.

This information tends to be included in papers in a table of baseline characteristics.

EDIT: My original post incorrectly stated that the interquartile range for the control group was 75-570. Thanks to @BruceInOz for pointing out the error - my eyes had slid down one row in the table! All other numbers in original post were accurate.

 

[Barry53]

It's also important to keep in mind that the 4.2 drop between subjects and controls is not a percentage. The paper never refers to it as a percentage of anything. It would be 12.7% of 33, but they never say that, either. It's just a drop in the mean score between subjects and controls. The precise meaning of that oddly precise difference in scores seems... unclear.

I seem to remember another case where the answer was something like 4.2 - and yet it was unclear what the question was... Hmmm.... what could that have been...?

With a little bit of memory aid from google the answer the computer was working towards was "The Ultimate Answer to Life, The Universe and Everything", but warned that when it found the answer no one would understand it. After much keeping on tenterhooks it eventually churned out the answer - 42! Even though very cryptic, no doubt much more rational than some of this other crud we are dealing with.

 

[Barry53]

The deadline is 2 weeks from publication which was Thursday, June 22 so that would mean either Wednesday, July 5 or Thursday, July 6. So not that much time.

The letters can't be too long: 250 words or less, with 5 or fewer references. So still plenty of time to write a letter.

Do the letters have to be from medically savvy people? Or can they be from anyone who feels they have a slightly different but important take on a flaw/oversight in the paper?

 

[Barry53]

Just seen this in The Lancet's correspondence information:-

All accepted letters are edited, and proofs will be sent out to authors before publication

How heavy handed is this likely to be I wonder. That of course is even if something is accepted.

 

[Barry53]

I presume this applies irrespective of whether a letter is accepted for publication:-

... the journal might invite replies from the authors of the original publication, or pass on letters to these authors

 

[Barry53]

More likely there was an outbreak of mass hysteria when they discovered the light bulb did not want to change.

Old but great joke! The version I heard was "How many psychiatrists does it take to change a light bulb?" Answer: Only one but only if the light bulb really wants to change.

 

[Valentijn]

Duration of illness at baseline in the GETSET trial is reported as 46 months for the GES group (that's a median rather than a mean), with an interquartile range of 23-114 months (aka most fell within that range) and 42 months for the control group, with an interquartile range of 75-570 months.

Very weird to use the median instead of the mean. Based on the differences between the interquartile ranges, doesn't it suggest that the mean duration was probably much longer in the control group? Maybe even by a factor of 3-4 times longer?