• 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.

PACE Trial and PACE Trial Protocol

oceanblue

Guest
Messages
1,383
Location
UK
Maybe you could calculate the SMC. I calculated the difference in 2F would be less than 5.5 (4mm) and in 2G around 6.9.

Somebody pointed out the following to me:
When 95% CI arms cross that can mean the p-value calculated for the difference between means will be non-significant. For example, Fig. 2B shows no overlap of CI arms at 52 weeks if you compare SMC to GET or CBT so we can conclude there is a significant difference. However, Fig 2F shows clearly that SMC upper CI arms overlap with those of GET and CBT lower CI arms. A similar pattern is also seen in Fig 2C and Fig 2G. Taking these four Figs into account, it decreased fatigue is statistically significant but there might be no significant difference in physical functioning measured.

If one looks at 2F and 2G, one seems to have enough information to say SMC isn't different from CBT and GET.

If SMC was slightly higher in 2F and 2G to 2E, it could be sufficient to make the differences no longer significant.

One caveat is that these are unadjusted data but that may not matter. I do believe their overall single p-values* look at whether individual differences are still significant at each point.

*there are four time-points and four measures at each point, resulting in lots of comparisons that might be different
I don't think that point about overlapping arms is right: according to your ref (and others I've seen) arms can overlap by half a length and still be statistically significant at 95%.

As I was trying to say in the earlier post, since SMC has a higher baseline than GET/CBT you can't directly compare the 52-week end points.

IC SMC data: 52.1-39.1=+13.1
Full cohort: 50.8-39.3=+11.6 (nb lower increase for SMC full cohort)

Having done a quite a bit of work to get this data, I wish the results had been more 'interesting', but this is the way they turned out.:(

Intersting that the effect with fatigue appears to be quite a bit stronger than for physical functioning, which is seen in the full cohort too.
One caveat is that these are unadjusted data but that may not matter
Not sure how much difference adjusting makes in practice. The difference between CBT/GET and SMC I calculated from the graph data almost exactly matched the adjusted difference figure in table 3.
 

Dolphin

Senior Member
Messages
17,567
That point about overlapping arms is wrong, I think: according to your ref (and others I've seen) arms can overlap by half a length and still be statistically significant at 95%.
Have you seen any that overlap by a full arm and are not statistically significant. That paper is only talking about half an arm. It seems very likely to me trying to think about in terms of probability density functions with different frequencies at each point, that it would be significant in the case of 2F and 2G (if it was simply left at the final value i.e. ignoring any adjustments for the moment).

If one takes the example of two distributions with confidence intervals (CIs) of the same magnitude where the touch each others means: one already has half the area one needs as 50% of the time for the lower one, it will be above the end of the other arm (which will occur with a probability of 0.025 by definition i.e. 0.0125. One only needs it that more than 0.025 of the time, the lower one is higher than the upper one, and the confidence interval for the difference overlaps (at p=0.05).

For what it's worth, looking at Table 3, the unadjusted changes are +19.2, +20.0 and +11.6 for CBT, GET and SMC. This give unadjusted change differences of +7.6 (CBT vs SMC) and +8.4 (GET vs SMC); the adjusted change differences actually were: +7.1 (CBT vs SMC) and +9.4 (GET vs SMC).

IC SMC data: 52.1-39.1=+13.1
Full cohort: 50.8-39.3=+11.6 (nb lower increase for SMC full cohort)

Having done a quite a bit of work to get this data, I wish the results had been more 'interesting', but this is the way they turned out.
Not sure why you are disappointed. That 1.3/1.5 could make the difference and it may no longer no statistically significant.

The confidence intervals for the differences will be bigger than those in Table 3 because of smaller sample sizes as I understand it.

As I was trying to say, since SMC has a higher baseline than GET/CBT you can't directly compare the 52-week end points.
I missed what you were saying there; however adjustment for baseline values doesn't always mean that one simply subtracts - sometimes it can be more or less e.g. if values tend to plateau, those on 38 and 39 could end up with similar final values and one wouldn't adjust or the adjustment might be less. Basically, many times researchers don't look at change. Anyway, not a major point. The overlap is my major point.
 

Sean

Senior Member
Messages
7,378
We recommend that research studies use the SPHERE...

Now there is a warning bell. Anything that relies on that simplistic checklist assessment is going to be suspect.
 

oceanblue

Guest
Messages
1,383
Location
UK
Specialist Medical Care - includes quite a bit of pacing advice?

Below is the description of Specialist Medical Care taken from the Patient Information Leaflet given to patients before they decide to enter the trial - as detailed in the full protocol.

It looks to me like there is a significant element of Pacing in SMC that goes beyond simply saying 'don't do to much or rest too little'. Nb the SMC-only group had an average of 5 SMC sessions vs 3 for the other groups, so there was scope for quite a lot of advice. Specialist doctors included Simon Wessely too, who wouldn't just be there to write a prescription.

While the similarity between SMC and Pacing might have reduced any relative improvement with APT, I presonally think APT was doomed to failure by it's bizarre 'within 70% of perceived energy' limits so I'm not sure we can conclude anything from the trial re pacing.

Specialist Medical Care
Specialist medical care is the most usual treatment for CFS/ME, and it helps many people improve. You get a confirmed diagnosis, an explanation of why you are ill, and general advice about managing your illness. Your specialist might either prescribe medicine to help you manage troublesome symptoms such as insomnia and pain or advise your GP about what medicine is appropriate.

Here is some of the advice you may get as part of specialist medical care.
• Avoid extremes of activity. Many people with CFS/ME get into a pattern of being very
active and then very inactive. It is better to give yourself a pattern of activity that you
can keep going. This may be a lower level of activity you are used to.
• Set a daily level of activity. It will help to set a simple level of activity that you do every
day. Stretching exercises, for example, will minimise the weakening effects that creep
up if you don't use your muscles for a time.
• Make only gradual changes to your activity level. If you feel you can increase your level
of activity, and not everyone does, make changes carefully and gradually. A sudden
increase in activity may make your symptoms worse.
• Try to reduce stress in your life. When we are ill, stresses such as excessive work
demands don't help us. If you can reduce these stresses, it will help you recover.

Here's how the Lancet paper described SMC:
The manual was consistent with good medical
practice, as presently recommended.2 Treatment consisted of
an explanation of chronic fatigue syndrome, generic advice,
such as to avoid extremes of activity and rest
, specific
advice on self-help, according to the particular approach
chosen by the participant (if receiving SMC alone), and
symptomatic pharmacotherapy (especially for insomnia,
pain, and mood).
 

oceanblue

Guest
Messages
1,383
Location
UK
Recovery rates according to the original protocol - estimates

Since the results for recovery according to the original protocol (SF36 of 85 or more, bimodal chalder of 3 or less) will probably never be released I've had a stab at estimating them from the data we do have.

I've explained the logic below, but here are the results, with the PACE figures for 'within the normal range' in brackets

CBT: 12% 'recovered' (vs 30% 'within normal range')
GET: 12% (28%)
SMC: 5% (16%)

The authors might have commented:
With CBT or GET just over 10% of patients recover after a year, compared with 5% for the controls. We think we've licked CFS.
OK, they didn't say this, but they have said the therapies are moderately effective.

And now for the rationale. There's no substiture for having the correct calculations done on the real data but I don't think this will be a million miles out.
1. Assume SF36 and fatigue scores are normally distributed for groups at 52 weeks
Not an ideal assumption, but it is one that the authors make. Also, while data for a general population won't be normally distributed (data is heavily skewed towards everyone getting full-health scores, for a group of sick people (as these are even after therapy) the results will approximate more closely to a normal distribution.

Assuming a normal distribution, we can use group means/SDs (via Z-scores) to estimate the proportion of group members above any particular number, in this case the threshold scores.

2. Use SF36 results to estimate those meeting SF36 recovery threshold
Lets start by estimating the proportion of particpants with SF36 scores over the recovery threshold, using the data for CBT, GET and SMC at 52 weeks:

Therapy, mean, SD, Z-score, % over threshold:
CBT 58.2 24.1 0.905 18%
GET 57.7 26.5 0.842 20%
SMC 50.8 24.7 1.182 12%

3. Estimate the proportion of those exceeding SF36 threshold who also exceed the fatigue threshold
This is where it gets more uncertain.

Obviously not all those who are over the SF36 threshold will also exceed the fatigue threshold required for recovery. For the 'clinically useful difference threshold', about 80% of those who exceeded the SF36 threshold of 8 points also exceeded the fatigue threshold of 2 points. However, in the rarefied heights of recovery, it's likely that far fewer of those who meet one threshold also meet the other one. But what figures to use?

My approach was to do the same exercise for the 'back to normal' group since we know how many of those definitely met both thresholds. (The authors give the proportion of paticipants who exceed both fatigue and SF36 'normal' thresholds.) So I took the SF36 and estimated how many would meet the back to normal threshold of 60, then used an adjustment fudge factor to convert from my SF36 estimate to the given figures for both thresholds:

Therapy, Est SF36 threshold, [Known % over both thresholds] = adjustment factor
CBT: 47% [30%] = 64%
GET: 47% [28%] = 60%
SMC: 36% [16%] = 45%
Would have been nice if the SMC fudge factor worked out closer to CBT/GET.

Then I applied these fudge factors to the estimate above for those meeting SF36 recovery threshold alone - et voila!

OK, far from perfect but I'm open to suggestions for better methods. Summary:
  1. Use mean and SD to estimate proportion over SF36 recovery threshold
  2. Estimate proportion of those exceeding SF36 threshold who also meet fatigue threshold
  3. = proportion recovered
 

Dolphin

Senior Member
Messages
17,567
oceanblue, interesting you're looking at it.
For SF-36, you used 80. It should be 85 (or 82.5???). Thinking about it, for CFQ, they had a figure like 18.8 and said the threshold was the next lowest number (18) so 85 may seem reasonable in such circumstances i.e. the next possible number.
Maybe you could calculate the figures again. I'm not sure on that. Maybe you could calculate it both ways (82.5 and 85).

Also, here's the full "recovery" definition:
"Recovery" will be defined by meeting all four of the following criteria: (i) a Chalder Fatigue Questionnaire score of 3 or less [27], (ii) SF-36 physical Function score of 85 or above [47,48], (iii) a CGI score of 1 [45], and (iv) the participant no longer meets Oxford criteria for CFS [2], CDC criteria for CFS [1] or the London criteria for ME [40].

If you wanted to bring in CGI info, you could consider using Kathy Fulcher's PhD http://hdl.handle.net/2134/6777 which has raw data on SF-36, CGI, etc. for 66 or 33 exercise and 33 flexibility then exercise (although there is a bit of missing info). For the first 33, one needs CGI2full as the others are just rated 1/2 (improver/non improver). So for example number 6 got to 95 but only had a CGI of 2. One even has Chalder Fatigue Scale scores (called Wessely and Wesp and Wesm for physical fatigue and mental fatigue scores although unfortunately this is the 14 question version. But 18*(14/11)=22.91 so you could use the threshold of 23 (say). I'm not necessarily looking for a lot of work - just the breakdown of the CGI scores of those scoring >=85 on SF-36 and <=23 on "Wessely fatigue".
ETA: number #6 has a Wessely fatigue score of 20 (out of 42) i.e. with a SF-36 PF score of 95 and a Wessely fatigue score of 20 (out of 42) (almost certainly <=18 out of 33), this person didn't have a CGI score of 1 and so wouldn't be counted in "recovery".
 

oceanblue

Guest
Messages
1,383
Location
UK
oceanblue, interesting you're looking at it.
For SF-36, you used 80. It should be 85 (or 82.5???). Thinking about it, for CFQ, they had a figure like 18.8 and said the threshold was the next lowest number (18) so 85 may seem reasonable in such circumstances i.e. the next possible number.
Maybe you could calculate the figures again. I'm not sure on that. Maybe you could calculate it both ways (82.5 and 85).
Good point, thanks. Recalc on 85 gives estimated recovery rates as defined by original protocol of:
CBT & GET: 9%
SMC: 4%


that's probably a better estimate.
Also, here's the full "recovery" definition:
Basically I assumed if you hit both sf36 and cfq thresholds you would almost certainly score yourself 'very much better' and very likely no longer hit any CFS criteria but, yes, the figures above might need a little shaving off them. however, the margin of error of my approach is quite wide already so I'm not sure I want to finesse this much more; it might not add much to accuracy, esp as I've already used 85 instead of 82.5 on sf36.

Anyway, I think that looks better now; feel free to add you own assumptions to this and I can supply my spreadsheet if that helps.
 

Dolphin

Senior Member
Messages
17,567
Good point, thanks. Recalc on 85 gives
CBT & GET: 9%
SMC: 4%

that's probably a better estimate.

Basically I assumed if you hit both sf36 and cfq thresholds you would almost certainly score yourself 'very much better' and very likely no longer hit any CFS criteria but, yes, the figures above might need a little shaving off them. however, the margin of error of my approach is quite wide already so I'm not sure I want to finesse this much more; it might not add much to accuracy, esp as I've already used 85 instead of 82.5 on sf36.

Anyway, I think that looks better now; feel free to add you own assumptions to this and I can supply my spreadsheet if that helps.
Thanks for that.
Figures of under 10% (or 5% more than SMC) could be quite easily dismissed as not very impressive at all esp. when it's just an Oxford cohort to start with, relatively short illness duration on average (they could tell us the characteristics of these people), lots of people opted not to take part, didn't have any co-morbidities that would exclude them from taking part, these figures don't necessarily represent full recovery (many of the people may still not be working so we don't know how they might cope trying to work), etc.

Yes, feel free to send on Excel file, thanks.
 

WillowJ

คภภเє ɠรค๓թєl
Messages
4,940
Location
WA, USA
Isn't that the crux.

CBT and GET groups are told repeatedly that they will get better. SMT and APT groups are not.

Any wonder the CBT and GET groups do better on the subjective self report 'happy sheets' and a case could certainly be made that the CBT and GET arms both received 'non standard encouragement' prior to the 6MWT.

@all:
I thought the APT group was told pacing would create the most likely conditions for "natural recovery"?

How much difference is there between the CBT and GET ("this will cure you"?) and the APT "this is not a cure but will create the best conditions for a possible natural recovery" themes?
 

urbantravels

disjecta membra
Messages
1,333
Location
Los Angeles, CA
I was the one who raised the issue of expectations management and the nocebo effect vs. placebo effect, especially when outcomes are being measured by subjective self reports.

It's very clear that in PACE, there was not even an attempt at objectivity in how the different treatments were presented to the patients. The CBT and GET arms were told that they could get better and that the therapy would help them do so; the APT people were told that there was nothing that could be done to cure their disease, but that by living within their limits they "might" create the conditions for "natural recovery" to occur. (We all know how good it feels waiting around for that to happen...), versus being told repeatedly that the therapy being given is going to improve you, and that you can improve your health via your own efforts. There was no comparable "Is this a cure? Be honest, the answer is no" statement in the CBT and GET manuals, as there was in the APT manual.

I really don't understand why this alone doesn't get the whole study discredited on ethical grounds.
 

Angela Kennedy

Senior Member
Messages
1,026
Location
Essex, UK
I was the one who raised the issue of expectations management and the nocebo effect vs. placebo effect, especially when outcomes are being measured by subjective self reports.

It's very clear that in PACE, there was not even an attempt at objectivity in how the different treatments were presented to the patients. The CBT and GET arms were told that they could get better and that the therapy would help them do so; the APT people were told that there was nothing that could be done to cure their disease, but that by living within their limits they "might" create the conditions for "natural recovery" to occur. (We all know how good it feels waiting around for that to happen...), versus being told repeatedly that the therapy being given is going to improve you, and that you can improve your health via your own efforts. There was no comparable "Is this a cure? Be honest, the answer is no" statement in the CBT and GET manuals, as there was in the APT manual.

I really don't understand why this alone doesn't get the whole study discredited on ethical grounds.

Yes- you're right I think, Urban. Malcolm Hooper of course relates to the situation in 'Magical Medicine'. It is quite shocking.
 

Dolphin

Senior Member
Messages
17,567
If anyone wants to send in a reply to this, it'd be appreciated.
It was included in a free newspaper for Irish doctors.


Last year, following publishing a piece on the Santhouse et al. editorial in the British Medical Journal, they published not one but five letters over a series of weeks (John Greensmith, Tom Kindlon, Gerwyn Morris, Orla Ni Chomhrai & Vance Spence (only two with Irish addresses) - that was most of the people who wrote in, as I recall.
They may be glad to fill up space in their newspaper.


People can also put comments online but letters would be preferred. You can always post your letter as a comment if you prefer.


Probably best to keep letters under 400 words and ideally less than that again.
Address is: editor@imt.ie that's editor @ imt.ie


Don't forget to put your address in the letter and also a telephone number (which won't be published).


Thanks

http://bit.ly/hAvLon
i.e.
http://www.imt.ie/clinical/2011/03/cognitive-behavioural-therapy-not-harmful-in-chronic-fatigue.html
They published one letter this week:
http://www.imt.ie/opinion/2011/03/caveat-needed-on-safe-and-effective-declaration.html

Don't forget - they published 5 over a number of weeks last year so plenty of time to write in.
 

Dolphin

Senior Member
Messages
17,567
Just came across the following.
The point I'm highlighting is the only relevant one: the population that was used for the norms for the Chalder Fatigue Questionnaire was people who attended their GP in the last year. If these figures were correct, then a large chunk of the population were missed. I think oceanblue mentioned 80% of the population go to their GP every year earlier in the thread (??). Anyone else ever come across a figure (esp. for the UK).

http://www.gponline.com/News/article/1043685/Demeaning-phrases-stop-men-checking-health/
Demeaning phrases stop men checking health By Tom Moberly, 29 November 2010

Demeaning phrases used to describe male illness are putting men off accessing health services, a survey suggests.

Demeaning phrases have such as 'manflu' prevented men from seeing a GP

In a survey of 3,000 people, 52% of men said the term 'man flu' had prevented them seeking advice for a legitimate illness.

In addition, 53% of men said they worried about wasting GPs' time. Only 55% of men visit their GP once a year or more, compared with 72% of women.

Many men also said they would not visit their GP if they suffered persistent thirst (80%), frequent urination (77%) and erectile dysfunction (75%), even though such symptoms could be early warning signs of underlying health conditions.

The survey included 1,500 men and was conducted on behalf of Pfizer.
 

Dolphin

Senior Member
Messages
17,567
If anyone wants to send in a reply to this, it'd be appreciated.
It was included in a free newspaper for Irish doctors.


Last year, following publishing a piece on the Santhouse et al. editorial in the British Medical Journal, they published not one but five letters over a series of weeks (John Greensmith, Tom Kindlon, Gerwyn Morris, Orla Ni Chomhrai & Vance Spence (only two with Irish addresses) - that was most of the people who wrote in, as I recall.
They may be glad to fill up space in their newspaper.


People can also put comments online but letters would be preferred. You can always post your letter as a comment if you prefer.


Probably best to keep letters under 400 words and ideally less than that again.
Address is: editor@imt.ie that's editor @ imt.ie


Don't forget to put your address in the letter and also a telephone number (which won't be published).


Thanks

http://bit.ly/hAvLon
i.e.
http://www.imt.ie/clinical/2011/03/cognitive-behavioural-therapy-not-harmful-in-chronic-fatigue.html
They published one letter this week:
http://www.imt.ie/opinion/2011/03/caveat-needed-on-safe-and-effective-declaration.html

Don't forget - they published 5 over a number of weeks last year so plenty of time to write in.
They published 2 more today i.e. 3 this week:
http://www.imt.ie/opinion/2011/04/p...for-the-‘biggest-rift’-among-me-patients.html

http://www.imt.ie/opinion/2011/04/the-pace-study-is-out-of-step-with-mecfs.html
 

Dolphin

Senior Member
Messages
17,567
What on earth is "man flu"? Some slang term that hasn't made it to the US yet, apparently.
It's a term that the media have used occasionally and the public now use sometimes - it involves claiming men if they have a sniffle say they have influenza/make a big deal out of their minor symptoms.
 

urbantravels

disjecta membra
Messages
1,333
Location
Los Angeles, CA
I'm so confused now!! I thought it was women that were hypochondriacs, exaggerating minor symptoms, making a big fuss and pestering doctors unnecessarily. Now it's men too?

Clearly the problem is more widespread than we thought - it's sick people that are unreliable, irrational exaggerators who need to shut up and not complain so much about being "sick" or talk so much about their "symptoms." Got you.
 

WillowJ

คภภเє ɠรค๓թєl
Messages
4,940
Location
WA, USA
yes, because doctors can find out everything they need to know via a cursory visual examination and standard screening tests. they do not want or need patients to "self-report" the "status" of being "sick" or experiencing "symptoms"
 

Dolphin

Senior Member
Messages
17,567
yes, because doctors can find out everything they need to know via a cursory visual examination and standard screening tests. they do not want or need patients to "self-report" the "status" of being "sick" or experiencing "symptoms"
Except when they are doing UK5m (US$8m) trials of CBT and GET in which case what patients self-report on particular questionnaires is perfectly fine.