PACE Trial and PACE Trial Protocol

[Dolphin]

I've started a new thread on the Wessely papers from the 90s, and have replied there re the graph:

Could you explain a bit more about diabetes and behaviour?

It would seem to me that there is plenty of potential that managing one's diabetes better could lead to better glycaemic control/lowering of HbA1c in some people which could improve scores on average. For example, I have a distance relative, a widower, who prays every night he won't wake up - this was even before he had many complications from the diabetes. He has bags of sweets in his house and generally doesn't look after himself as a diabetic should.

I'm no expert on diabetes but some or all of these would seem to be true:
(i) diet can play a big part in controlling it. There is some foods (and drinks) one should generally avoid or only take in very small quantities. If a person wasn't doing this well, a motivational program (which Simon Wessely said was involved in intervention) would likely improve HbA1c levels.

(ii) more exercise can stabilise blood sugar levels as far as I know (at least in type 2 but I think I've seen also in type 1). There would be plenty of people with diabetes who for one reason or another haven't prioritised this/aren't doing enough/could do more exercise.

(iii) measuring blood sugar levels - some people may not do this sufficiently often to calculate when to eat/how much to eat and when to take insulin. Related to this point is whether people take their insulin at the optimum times.

(iv) adherence to other drugs that might help with blood sugar control.

So to me, I can see how a management program or if necessary a CBT program if somebody wasn't complying well enough, could on average improve scores in diabetes.

Such a program could be indirect also. For example, taking the example of my depressed distant relative: if his depression was treated, he might be more motivated to follow diabetes advice.

In ME/CFS, I think it is much less clear cut what we should be recommending people. In particular, I'm not convinced about the model underlying GET and GET-based CBT.

So just because a CBT intervention might help on average in diabetes, doesn't tell us that much about what a GET-based CBT intervention (for example) will do in ME/CFS.

 

[oceanblue]

It would seem to me that there is plenty of potential that managing one's diabetes better could lead to better glycaemic control/lowering of HbA1c in some people which could improve scores on average.

I'm no expert on diabetes but some or all of these would seem to be true:
(i) diet can play a big part in controlling it. There is some foods (and drinks) one should generally avoid or only take in very small quantities. If a person wasn't doing this well, a motivational program (which Simon Wessely said was involved in intervention) would likely improve HbA1c levels.
...

So to me, I can see how a management program or if necessary a CBT program if somebody wasn't complying well enough, could on average improve scores in diabetes.

Such a program could be indirect also. For example, taking the example of my depressed distant relative: if his depression was treated, he might be more motivated to follow diabetes advice.

In ME/CFS, I think it is much less clear cut what we should be recommending people. In particular, I'm not convinced about the model underlying GET and GET-based CBT.

So just because a CBT intervention might help on average in diabetes, doesn't tell us that much about what a GET-based CBT intervention (for example) will do in ME/CFS.

Oh, I see, thanks. CBT seems to be used to increase adherence to a self-management programme. It's not, as Simon Wessely implied, that changing patients beliefs about their illness magically improves the underlying physiology of things, it's just they get better at doing sensible things with diet/medication/self-care which directly impact on blood sugar levels. I mean, it's a good result if it improves people's health but says nothing about the ability of CBT to fix 'medically unexplained' problems. For someone who champions the idea of 'evidence-based medicine' he really does play fast and loose with the evidence.

 

[Dolphin]

4.Data from the original study indicate this is an unhealthy cohort.
According to Pawlikowska, 38% of patients had a score about the original Chalder bimodal cut off of 3 (as used in the PACE protocol) and 18.3% of patients were substantially fatigues for 6 months or longer. Whoa, that looks unhealthy, esp as the paper quotes a 1990 paper that found only 10% of GP practice patients had fatigue for one month or more. I think there are some US studies indicating fatigue of over 6 months in the population is much less than 18%.

This CDC study:

Arch Intern Med. 2003 Jul 14;163(13):1530-6.
Prevalence and incidence of chronic fatigue syndrome in Wichita, Kansas.
Reyes M, Nisenbaum R, Hoaglin DC, Unger ER, Emmons C, Randall B, Stewart JA, Abbey S, Jones JF, Gantz N, Minden S, Reeves WC.

http://archinte.ama-assn.org/cgi/reprint/163/13/1530
gives a weighted* prevalence for Respondents fatigued for 1+ month: a prevalence of 10.84% (=29566/272,838) i.e. a lot lower than the Pawlikowska for for 6+ months.

* i.e. adjusting for sample - this is a good one to have

ETA:

Abstract

Previous estimates of the
prevalence of fatigue and chronic
fatigue have derived largely from
treated populations and have been
biased by differential access to
health-care treatment linked with
gender, racial/ethnic and social
class status. This study involves a
community-based prevalence
study of prolonged fatigue and
chronic fatigue. It addresses: (1)
the rate of prolonged fatigue and
chronic fatigue in a socioeconomically
and ethnically
diverse sample of 28,673 adults
in Chicago; and (2) establishes
the relative prevalence of
prolonged fatigue and chronic
fatigue across race/ethnicity,
socio-economic status and
gender. Univariate and
multivariate statistical techniques
were utilized to delineate the
overall rate of prolonged fatigue
and chronic fatigue in the
Chicago population and its
relative prevalence by gender,
race/ethnicity, and social class.
Findings indicated that fatigue is
common in urban populations,
but that prolonged fatigue and
chronic fatigue occur in about
5.00 to 7.68 percent and 2.72 to
4.17 percent, respectively, of the
sample of the population. Highest
levels of fatigue were consistently
found among women and those
with lower levels of education
and occupational status.

11.Jason LA, Jordan KM, Richman JA, Rademaker AW, Huang CF, McCready W, Shlaes J, King CP, Landis D, Torres S, Haney-Davis T, Frankenberry EL. A community-based study of prolonged fatigue and chronic fatigue. J Health Psychol 1999; 4: 926.

 

[oceanblue]

Dolphin, thanks for digging out those studies, I can't even find stuff on my own computer now.

To Recap from all these studies:

One month+ fatigue rate:
David 1990: 10%
Jason 1999: 5.0% - 7.7%
CDC (Reyes 2003): 10.8%
Pawlikowska 1994: 38%

6 month+ fatigue rate:
Jason 1999: 2.7% - 4.2%
Pawlikowska 1994: 18%

Jason & Reyes studies are population studies so the most reliable.

And of course, the Cella cohort used as 'normative' by PACE are slightly more fatigued than the Pawlikowska cohort.

 

[Dolphin]

Dolphin, thanks for digging out those studies, I can't even find stuff on my own computer now.

Aside: I used to do a lot of searches for terms in my folder that has all the research paper; however after an "upgrade" of (free) Acrobat Adobe reader, it no longer searches the contents of pdfs which is frustrating.

 

[Snow Leopard]

The Diabetes type 1 CBT/motivational interventional is not really analagous to ME/CFS: with diabetes, it is very clear what are the sort of behaviours to encourage; it is much less clear cut in ME/CFS.

The Diabetes study results reflect the same sort of dishonesty we've come to expect from UK psychiatrists. Those wonderful results? Well guess what - the 95% CI overlaps.
http://www.annals.org/content/149/10/708.figures-only
http://www.annals.org/content/149/10/708.full.pdf+html

The secondary measures were also statistically insignificant.

If this was a drug therapy, it wouldn't get approval.

 

[oceanblue]

The Diabetes study results reflect the same sort of dishonesty we've come to expect from UK psychiatrists. Those wonderful results? Well guess what - the 95% CI overlaps.
http://www.annals.org/content/149/10/708.figures-only
http://www.annals.org/content/149/10/708.full.pdf+html

The secondary measures were also statistically insignificant.

If this was a drug therapy, it wouldn't get approval.

My understanding is that 95% error bars can overlap up to halfway and the difference in means will still be significant, so motivation vs std care is not significant but motivatin+cbt vs std care is just significant - but probably not clinically important. This from the text:

The estimated differences in time-averaged treatment
effects of change in hemoglobin A1c levels were -0.07%
(95% CI, -0.31% to 0.16% ie not significant) for motivational enhancement
therapy versus usual care and -0.25% (CI, -0.49%
to -0.01% just significant) for motivational enhancement therapy plus
cognitive behavior therapy versus usual care (Appendix Table,
available at www.annals.org).

Also, cbt+motivation was not significantly better than motivation alone. Hard to see the point of cbt, then. Oh, and although they gave the cost of cbt as 80 a session, they didn't do the sums to for a whole course: about 1,100 for a piddly benefit, if you're interested.

Anyway, it's hardly the triumph that Simon Wessely implied.

 

[Snow Leopard]

I think the 'rule of thumb' is 25% of the bars - but in that case, the only significant result was at 12 months and not at 6, 9 months which is rather strange. It is difficult to conclude that the therapy was beneficial from these results.

Edit - note, "25% of the bar" = "50% of one arm", corresponding to approximately p=0.05.

 

[Angela Kennedy]

Oh, I see, thanks. CBT seems to be used to increase adherence to a self-management programme. It's not, as Simon Wessely implied, that changing patients beliefs about their illness magically improves the underlying physiology of things, it's just they get better at doing sensible things with diet/medication/self-care which directly impact on blood sugar levels. I mean, it's a good result if it improves people's health but says nothing about the ability of CBT to fix 'medically unexplained' problems. For someone who champions the idea of 'evidence-based medicine' he really does play fast and loose with the evidence.

And even in this diabetes scenario- the above example might benefit more from supportive counselling therapies (If they guy has issues around not wanting to wake up- that sounds like psychological distress to me, and that might be why he's not following the correct diet either.)

 

[oceanblue]

I think the 'rule of thumb' is 25% of the bars - but in that case, the only significant result was at 12 months and not at 6, 9 months which is rather strange. It is difficult to conclude that the therapy was beneficial from these results.

Yes, I've seen different rules of thumbs, probably best to use the actual data and confidence limits (Table 2). The text I quote above is for the time-weighted average (-0.25% (CI, -0.49% to -0.01%)) rather than any one time point, which is presumably the most relevant measure. So I think the statistically significant claim is valid (just), though probably not a useful benefit.

 

[Dolphin]

Eyeballing CIs in figures

Hope123 sent me before this info on reading overlap bars:

About 95% CI arm overlap: http://www.cscu.cornell.edu/news/statnews/stnews73.pdf
How to eyeball CIs in figures - interesting! : http://www.ncbi.nlm.nih.gov/pubmed/18991332

 

[Snow Leopard]

Thanks Dolphin.

Edit - nevermind, not really concerned with the specific statistics question I had.
If the time averaged results were from 3-9 months for the usual care (to exclude the bump) and 6-12 months for CBT, I'm not convinced that there would be a difference with p<0.05.

 

[oceanblue]

Leaving aside the issue of statistical significance, I don't think these results are of clinical importance anyway. To put the final HbA1c levels achieved with CBT, of about 9%, into context here's some info about HbA1c levels, http://www.diabetes.org.uk/hba1c

For non-diabetics, the usual reading is 3.5-5.5%. For people with diabetes, an HbA1c level of 6.5% is considered good control, although some people may prefer their numbers to be closer to that of non-diabetics.

People at greater risk of hypoglycaemia, may be given a target HbA1c of 7.5% to prevent too many low blood sugars from occurring.

So the time-weighted average improvement of 0.25%, giving a final level of about 9%, still leaves patients well over the target level for those at greatest risk of hypoglycaemia and the gain is only a very small step towards that 7.5% target (let alone the 6.5% target that represents 'good control').

 

[Sean]

So the time-weighted average improvement of 0.25%, giving a final level of about 9%, still leaves patients well over the target level for those at greatest risk of hypoglycaemia and the gain is only a very small step towards that 7.5% target (let alone the 6.5% target that represents 'good control').

Any biophysical therapy that delivered such small gains would be ignored, except maybe as a possible source of not particularly promising research ideas.

 

[Dolphin]

Lancet correspondence has now been published

Link to Lancet: http://bit.ly/j9cEoD or http://www.thelancet.com/search/res...Name=AllFields&journalFromWhichSearchStarted=

PACE Trial - letters that were published and authors' response:
http://forums.phoenixrising.me/show...published-and-authors-response-(and-editorial)

"Patients' power and PACE" (Lancet editorial accompanying PACE Trial correspondence)
http://forums.phoenixrising.me/show...torial-accompanying-PACE-Trial-correspondence)

 

[Snow Leopard]

Leaving aside the issue of statistical significance, I don't think these results are of clinical importance anyway. To put the final HbA1c levels achieved with CBT, of about 9%, into context here's some info about HbA1c levels, http://www.diabetes.org.uk/hba1c

So the time-weighted average improvement of 0.25%, giving a final level of about 9%, still leaves patients well over the target level for those at greatest risk of hypoglycaemia and the gain is only a very small step towards that 7.5% target (let alone the 6.5% target that represents 'good control').

I agree, thanks.

 

[Dolphin]

PACE Trial data - Freedom of Information request (MRC)

I have just been made aware of the following ongoing Freedom of Information request regarding the PACE Trial data: http://www.whatdotheyknow.com/request/request_for_raw_data_from_the_pa#outgoing-127757

I thought it might be better to start a separate thread on this in case anyone was interested. It's at: http://forums.phoenixrising.me/show...rial-data-Freedom-of-Information-request-(MRC)

 

[Dolphin]

AfME Interaction Piece on the PACE Trial

AfME Interaction Piece on the PACE Trial:
http://www.afme.org.uk/res/img/resources/PACEinIA76.pdf

I found it disappointing. Derek Pheby made a few good general points but there were lots more that weren't made - I think the regular contributors to this thread could do a lot better/add a lot more.

 

[oceanblue]

ME Association and others request PACE data according to protocol

From the MEA website:

PACE Trial text of joint letter calling for more information, 24 May 2011
by tonybritton on May 24, 2011

The ME Association together with the Young ME Sufferers Trust and the West Midlands ME Groups Consortium are calling on the PACE Trial organisers for some more information. We sent the following joint letter to one of the principal investigators, Professor Peter D White, on 24 May 2011.

 

[Bob]

----- Original Message -----
From: Mullan, Zoe (ELS-CAM)
To: ANGELA KENNEDY
Sent: Friday, March 04, 2011 12:18 PM
Subject: RE: The PACE trial

Dear Angela,

My apologies for the delay in getting back to you. We were not aware of any objections to this study, and we of course made sure that the trial protocol had been approved by an ethics committee before we decided to proceed with it. The trial received unanimous support from three clinical reviewers and a statistician. On this basis, we do not see a reason to deviate from the usual course of practice and make the peer review documentation public.

If you believe the journal has acted inappropriately over the publication of this trial, you should put your concerns in writing to our independent Ombudsman, Dr Charles Warlow, who can be contacted via ombudsman@lancet.com.
Best wishes,
Zo Mullan
Senior Editor
The Lancet
32 Jamestown Road
London NW1 7BY
UK
(T) +44 (0)20 7424 4910
(F) +44 (0)1865 853016

Just thought I'd re-post this correspondence sent to Angela from the Lancet, in case people aren't aware that the Lancet has an 'independent' Ombudsman to make complaints to.