Michael Sharpe skewered by @JohntheJack on Twitter

The point that seems alien to MS, is that good science - really good science - effectively controls against bias. People inevitably have bias (they can't not have), which is why scientific papers have to be underpinned by facts and scientific rigour, not merely subjective opinions and lax science. Again MS misdirects by suggesting that because people have a bias, that inevitably renders their work to be biased; maybe he speaks from personal knowledge here. A key point of scientific rigour surely, is to avoid personal bias by relying on methods largely immune to bias. Nothing is perfect of course.

 

Everyone has bias, it's what you do or don't do with it that counts. Sticking to rigorously good science practices (both in trials and papers) seems a good way for those with integrity to help ensure any bias cannot filter into their work.

 

He's breaking the world speed record for going from discovering Twitter to mastering the art of trolling.

 

Lets not forget also the promised and delivered fast tracking of the PACE trial by the Lancet.

 

Actually human bias is just a small component. Confounds in cohort selection, for example, can severely alter outcomes if some factor is involved that affects the outcome and was not dealt with by randomization.

Systematic biases, like in using the Oxford criteria, are also a big issue.

The selection of outcome criteria is major too. If you measure the wrong things the study is worthless. Oh ...

PACE had one of the WORST biases build in. The use of SD on population data for SF36PF is not defined. Its problematic. It results in highly skewed results. PDW wrote a paper on this in 2007. With the specific data in PACE it also necessarily requires the rise of the Superhumans, those who have score over the maximum of 100. Such a blatant and deliberate abuse of standard deviation is deliberate manipulation of data. No question.

 

Better?

 

Well this is just an attempt to save his house of cards. If this was acceptable behaviour then drugs could be approved by not passing clinical trials but by researchers saying they liked it and the FDA saying this is good enough to approve it...

I like this analogy, so patriarchal, so empty, so ridiculous except to those who believe in treating people badly or don't know enough to not fall for it

That is probably the easiest for the observer to understand talking point, we know the technical stuff but we can't expect a laymen to, but this is so ridiculous that it makes the point for anyone who hears it. Of course the reality deniers who wrote it and their apologists won't accept any proof of their malfeasance, they can't otherwise they are in lots of hot water...

 

I'm not sure there's much point trying to second guess what he is saying. Much of it is post hoc. And besides, reducing one of the criteria (PF) to well below the criteria set for improvement (>75 or >70) doesn't make any sense at all, even if combined with other measures.

Also, the KCL website summary on the whole recovery thing is very vague - PW is quoted as saying: "We focussed on recovery from the current episode of the illness", although how you can possible tell that from the data they used, I have no clue. In fact, all the quotes on that page say slightly different things.

 

MS: "But a regulated trial makes it very hard to bias the results."

That simply isn't true. But it is quite baffling that no-one on the steering committee spotted that they were overseeing an unblinded uncontrolled study with subjective composite endpoints. Bias galore!

Oh, but psych studies have been doing it that way for decades! Well, yeah... and... errr... yeah!

The main problem they have is that they simply don't understand the measures they have invented (CFQ). They don't know how it works, what it says, how to interpret it, how it varies within and between participants (I could go on all day...). They then don't understand what happens to it when you interfere with it (and PF) using the interventions. Subjective measures are not stable - which is why you need associated objective measures to make sure things stay on track (and to detect any bias). Any divergence means you shouldn't be using them any more. Instead, they interpreted that to mean that it was the objective measures that should be chucked. *facepalm*

 

Ye kind of like claiming a headache can be cured by hands on faith healing cos after the fifth day of the headache it went after faith healing for the first four days.

 

A single patient means nothing to him, as long as science doesn't turn on his miraculous 'treatment'. Thats the whole point of these games, a desperate and doomed to failure ploy to prevent anyone important from realizing the emperor has no clothes and funding real research.

 

Which of course is absurdly unscientific, implicitly admitting they did not focus on recovery in any real sense of the word at all! In this statement PW was admitting failure to achieve any kind of real recovery, effectively stating that more episodes would follow. Every time they skirt around one truth, they open up another hole to drop themselves into. One day, when their chickens come home to roost, there will be so much of this sort of material available.

 

If they show that CBT and GET can cause recovery on an episode-by-episode basis, that means sufferers have to keep going back every time they have an episode. An army of dependent sufferers for ever - how many jobs for the boys and what kind of an empire would that create?

 

As so many patients experience relapsing/remitting episodes anyway, then we're back to - you might as well have done nothing.

 

Isn't that just pacing anyway, to get maintain a low ceiling level in an incurable neurological disease. If its not pacing I am at a loss to understand what is.

 

"Bias is eveywhere".

But the wonderful scientific community are self correcting?

Interesting however when the FOI hearing happened Trudie Chalder contradicted this notion by admitting the Cochrane review wasn't independent because 3 of the PIs from PACE were on the review panel.

She had to do that in order to attempt to claim that they hadn't previously shared data with outside groups to attempt to defend themselves from not releasing data to other groups on request.

 

No, I don't think it is. It is symptom-denying; then the inevitable crash.

If there is any pacing involved, it is despite the therapy, not because of it.