PACE raw data available

trishrhymes

Senior Member
Messages
2,157
Likes
17,714
Anyone got any idea what the column headings mean?

trialarm cfqlsov0 cfqbsov0 pcfqls52 pcfqbs52 dgiq52F pfov0 p_pfov52 pgiq52F wtmts.0 wtmts.52 o_ov52cor

And which treatment arm (1 to 4) is which?

Edit - I clicked the corners and words magically appeared explaining them, not sure I'll find it again!
 

user9876

Senior Member
Messages
4,556
Likes
18,065
trialarm Which group people were in 1 GET, 2 APT, 3 SMC, 4 CBT

cfqlsov0 - Chalder Fatigue questionnaire -'likert' scores at baseline
cfqbsov0 - CFQ - binomial scores at baseline
pcfqls52 - (proforma) Chalder Fatigue questionnaire -'likert' scores at 52 weeks
pcfqbs52 - (proforma) Chalder Fatigue questionnaire -'binomial ' scores at 52 weeks
dgiq52F - Doctor rated CGI score at 52 weeks
pfov0 -- sf36 physical function score at baseline
p_pfov52 -- sf36 physical function score at 52 weeks
pgiq52F - Patient rated CGI score at 52 weeks
wtmts.0 - 6 minute walking test distance at 0 weeks
wtmts.52 - 6 minute walking test distance at 0 weeks
o_ov52cor - Whether the assessor thought someone met the Oxford criteria at 52 weeks
 

wdb

Senior Member
Messages
1,390
Likes
2,970
Location
London
Has anyone had a good look at this data yet ? I've had a bit of a look over it but did not spot anything particularly noteworthy.
 
Messages
435
Likes
366
Has anyone had a good look at this data yet ? I've had a bit of a look over it but did not spot anything particularly noteworthy.
I'll offer this as a tentative observation (because of uncertainty about the application of the Oxford criteria at 52 weeks), but it appears that the post-hoc changes to the recovery thresholds allowed one participant to "recover" having walked only 206 metres in six minutes, which is about 1.28 miles per hour.

At baseline they were able to walk 302 metres in the alloted time, so their walking capacity actually declined significantly over the course of the year.

Worryingly, the clinic doctor rated them as "very much better".
 
Last edited:

Jan

Senior Member
Messages
458
Likes
2,689
Location
Devon UK
[QUOTE="Sam Carter, post: 766977, member: 80

At baseline they were able to walk 302 metres in the alloted time, so their walking capacity actually declined significantly over the course of the year.

Worryingly, the clinic doctor rated them as "very much better".[/QUOTE]

FFS! Unbelievable!
 

alex3619

Senior Member
Messages
13,798
Likes
37,554
Location
Logan, Queensland, Australia
I'll offer this as a tentative observation (because of uncertainty about the application of the Oxford criteria at 52 weeks), but it appears that the post-hoc changes to the recovery thresholds allowed one participant to "recover" having walked only 206 metres in six minutes, which is about 1.28 miles per hour.

At baseline they were able to walk 302 metres in the alloted time, so their walking capacity actually declined significantly over the course of the year.

Worryingly, the clinic doctor rated them as "very much better".
If this is confirmed we should highlight it as often as possible as proof of how poor the recovery criteria are.
 

trishrhymes

Senior Member
Messages
2,157
Likes
17,714
I've had a bit of a look, and several things struck me

The scores on the physical capability questionnaire seem to vary wildly when compared with the distance walked, so, as has been mentioned , people could be classed as 'cured' yet able to walk comparatively short distances, and others who scored very low on questionnaire walked further.

Added to this, about a quarter of each group didn't do the second walk at all. And significant numbers in each group walked shorter distances after a year, yet said they had improved on questionnaire. If those who didn't manage to do the walk had been scored zero and included in the group averages, I suspect a very different picture would emerge. I haven't done that calculation.

I conclude from this that the data used to draw conclusions is not only invalid, which we already knew because of the problems with conducting the trial, but we're also so inconsistent as to be completely unreliable.

Since both validity and reliability of statistical data are prerequisites for statistical analyses to have any meaning statistically, let alone clinically, this invalidates the whole trial.

I think it's a text book case of why psychiatrists reliance on questionnaires to measure physical symptoms such as fatigue and physical functioning is completely bogus. They should have had the patients wearing actometers throughout the trial and done 2day cardiopulmonary exercise tests at the start and finish at the very least.

What a pile of old garbage.

And what a travesty of justice for the patients involved who have had their willingness to risk their own health to take part in a trial of untested methods completely screwed up for the sake of the careers of a few greedy, incompetent and totally unprincipled investigators. I won't call them scientists. This wasn't science.
 

A.B.

Senior Member
Messages
3,780
Likes
23,060
@wdb how did you deal with missing data? Often the data for the walking test at 52 weeks is missing.

I'm also looking at the data but haven't gotten past this yet:

=== CBT + SMC ===
Number of patients: 161
Patients that completed both walking tests: 122
Patients that only completed the walking test at 0 weeks: 37
Patients that only completed the walking test at 52 weeks: 1
Patients that didn't complete both: 1

=== GET + SMC ===
Number of patients: 160
Patients that completed both walking tests: 110
Patients that only completed the walking test at 0 weeks: 49
Patients that only completed the walking test at 52 weeks: 0
Patients that didn't complete both: 1

=== APT + SMC ===
Number of patients: 159
Patients that completed both walking tests: 110
Patients that only completed the walking test at 0 weeks: 44
Patients that only completed the walking test at 52 weeks: 1
Patients that didn't complete both: 4

=== SMC only ===
Number of patients: 160
Patients that completed both walking tests: 116
Patients that only completed the walking test at 0 weeks: 42
Patients that only completed the walking test at 52 weeks: 2
Patients that didn't complete both: 0
 

wdb

Senior Member
Messages
1,390
Likes
2,970
Location
London
@wdb how did you deal with missing data? Often the data for the walking test at 52 weeks is missing.
No point was plotted for cases with missing data.

This is interesting, can anyone verify,
SF36 at 52w mean for those who completed the 52w walk = 55.17318
SF36 at 52w mean for those who did not complete the 52w walk = 46.59766

if true suggests those who have missing data for the walk follow-up were significantly less fit and so would likely have brought the mean down had they completed the walk.

also

Doctor global impression mean
(completed walk) 2.668142
(missed walk) 3.388158

Patient global impression mean
(completed walk) 2.947939
(missed walk) 3.622378



(Please don't take as fact or quote anything posted here until it has been properly verified, we are not professional statisticians)
 
Last edited:

trishrhymes

Senior Member
Messages
2,157
Likes
17,714
I think one could justify the case for scoring all those who didn't do the walk as walking 0 metres, and including this in the averages. If they were unable to do the walk because they didn't feel well enough, they walked 0 metres. I suspect the researchers encouraged people not to do the walk if they weren't feeling up to walking for 6 minutes so they wouldn't register a low score and bring the averages down.

I'm intending for my own curiosity to have a look at how many in each group improved by 50 or 100 metres or more and how many went down by these figures, including those who didn't walk at all. But right now I have an appointment with my bed and last night's Strictly Come Dancing, then an hour or so lying still to recover!

I wonder whether it's a good idea to publish here analyses and graphs of our findings, or whether it's better to just discuss general points and hold off with the detail until, hopefully, Alem Mathees and the others working with him get their next more detailed paper published in a journal for the world to see.

I'm just a bit concerned that I, for one, might get something a bit wrong, or do a simplistic analysis that doesn't hold statistical water, and someone then spreads it around via twitter and we get hammered for being a bunch of amateurs.... I'd hate to give Wessely, White et al. any ammunition...