A general thread on the PACE trial!

Excellent thank you @Jonathan Edwards. I especially like the 'zero' slide. Epitomising how so little can say so much!

My wife and I were walking out in the countryside with our dog today. We don't get that far from the car, and we walk quite slowly, but get far enough to forget it's there for a little while. And we started chatting about how she feels better when doing this, and so without trying to lead her too much, I asked what it was that actually felt better, and what not. And when I got round to eventually asking "are you saying you feel less fatigued, less exhausted", she replied "oh no, that's worse, but I somehow have a better sense of well-being, compared to being stuck inside four walls most of the day". It is all so very subjective.

 

Excellent, thanks.

Can I ask what you mean by "if you want a placebo, CBT and GET may not be that good."

 

Sorry, I provided the explanatory link in my talk but left it to be inferred in the slides. The Rituximab study shows that a good going placebo response in ME roughly halves the gap between starting state and normality on a subjective scale. With the non-truncated Y axis the PACE figure is in comparison pathetic. That suggests that if, as Knoop and White claimed, CBT acts the way a placebo does then you may do better to be given a bag of saline with a magic label than spend weeks sitting with a therapist.

 

It's not the therapy!! It's the patient!!

https://www.ncbi.nlm.nih.gov/pubmed/15784798

 

The authors of that review don't seem to know what a "placebo response" even is. Changes in subjective outcome measures in a blinded control group from baseline cannot automatically be attributed to placebo effects as they can be due to a wide variety of biases.

 

The "magic label" was the hype and hope for Rituximab. It may have been the first time ever some patiens felt that effort was made to help and that they weren't ignored or belittled. This context cannot really be replicated.

CBT/GET probably doesn't work well as placebo because generally patients aren't convinced changes to thoughts and behaviour will do much.

 

Seems a bit difficult to compare the two as they used different ways to measure fatigue for example. If some of the items on the Chalder Fatigue Scale are largely irrelevant to ME/CFS patients condition, might that not cause improvements to look smaller?

I'm sure somebody suggested this before but wouldn't an RCT of CBT versus that bag of saline with a magic label on it, be a good way to convince others that there is no reliable evidence for GET/CBT?

 

There was a 4 cell RCT of immune therapy, CBT vs placebo respectively. No benefit was found.

Read Wessely et al comments to this study here:

https://sci-hub.se/https://www.ncbi.nlm.nih.gov/pubmed/7709961

 

Weird, considering it's well-established that expectation is as irrelevant as the nature of the placebo, or even being aware of being given a placebo, or how it's delivered or really any other factor.

And conventional wisdom? Jesus this guy is just too full of himself, doing the academic equivalent of "people are saying" because that's what he personally believes.

You can always count on Wessely to raise his personal musings to the level of groundbreaking insight, no matter how inane. This fool definitely gets high on his own farts.

 

PACE pretty much answered that question with the "usual care" comparison that showed no statistical difference that couldn't be accounted for by random chance or statistical artifacts.

That ship has sailed and sunk.

 

David Bell got a good response from patients using normal saline infusions which was thought to be because it restored blood volume. He said they could not do a clinical trial because how do you control saline and it was not useful as a treatment because of all the usual problems with infusions like infections at the site. One woman who was able to work as a lawyer while having weekly infusions carried on as long as she could but eventually had to give up.

The episode gave some insight into the disease but it has to be kept in mind if we are testing a drug infusion against saline controls. The difference in outcome between the two might not be as high because of the positive effects of the saline, and, of course, it may make a drug seem more effective if it is given in saline.

 

Technically that is true but the improvement with CBT and GET is very unimpressive by any measure.

 

Just reading this, I'm intrigued by the comment re: the ACR Improvement Criteria for rheumatoid arthritis. How could this be adapted for ME research? Would you sub in any other measures for ESR/CRP?

 

We had quite a long discussion about this at the other place about five years ago. I don't remember the details. The ACR system is a rather elegant compromise more than an idealised model - it follows the complex pragmatics of how consistent various measures are.

The basic idea though is to corroborate core subjective measures with objective findings using a multiple threshold scoring system. That means to score an improvement grade you need improvement both on core subjective feature and backup objective feature. Actometry could be the backup objective feature. Or like the ACR system you could have some alternative options. The only thing about allowing alternatives is that it increases the scope for fudging.

 

Thanks. I'll look up the other thread too. I agree it looks like a better solution.

 

Wessely/Deale/Chalder 1995 https://www.ncbi.nlm.nih.gov/pubmed/7709961

Which is totally contradictory to their conclusion from their review of "placebo" responses in clinical trials:
https://www.ncbi.nlm.nih.gov/pubmed/15784798

He also seems to be quite aware that patients don't actually increase their activity levels:

But this begs the questions:
(1) why do they assume activity is increased without actually measuring it directly?
(2) why do they assume an increase in activity without concurrent increases of symptoms is possible when no one has measured it?

They finish with this doozy:

Which has so far been a self-fulfilling prophecy, given the lack of research since.

 

I wonder why they were so confident in claiming "a condition with such a high placebo response as CFS"? Could it be an indication that the patients they were seeing, and developing their theories about, were unusual in some way?

 

Or is the same old cognitive mistake that many of us make? Assume your preconceptions are correct and try to cherry pick the facts to suit.

 

They have always been aware of the potentially strong placebo response from their therapies. It is discussed at some length after one of the presentations at the CIBA symposium.

 

Again, how does that explain this?
https://www.ncbi.nlm.nih.gov/pubmed/15784798