Michael Sharpe skewered by @JohntheJack on Twitter

You need to work on the sensitive discussion bit @Lucibee. The you will be an expert and can supervise others. They are always so sensitive with patient delusions.

 

I was only granted one question!

 

Godwin responds to Sharpe:
https://twitter.com/user/status/1008329868944306180

https://twitter.com/user/status/1008330673684348929

 

Great response it is.

 

Or wilfully blind...

 

I am pleased to see that Godwin is taking interest. I get what he is saying but it does seem a bit complicated. A simpler critique is that if you train one lot of people to try to believe they are improving and you do not train another lot then it tells you nothing useful if the first lot say they are improving and the others don't. Moreover, if the first lot show no increase in activity or work capacity you have reasonable evidence that the treatment does not do anything useful. Dr Sharpe seems to be lost even at that level.

 

"Any outcome is possible and must be respected"

Where PACE is concerned any outcome wasn't possible though. Original protocol outcomes weren't possible for a start once they'd made changes.

eta -plus other issues

 

Sharpe really does seem to be trying to persuade himself and everyone else that data entail interpretations. If you have a 'positive result' that means that you can predict that it will apply in routine practice. The whole point of complex trial design is that this is only the case if you are very careful about the conditions under which the result was obtained.

 

The 'model' is just 'GET and CBT are useful treatment for ME/CFS'.
PACE probably did not involve an adequate risk of showing this is wrong, if it is.
Although it does a pretty good job of showing it is wrong.

 

So why didn't they respect the outcome instead of changing the outcome measures to get the results they wanted?

 

It's great that Godwin is weighing in, although I don't really understand why in PACE the model was not falsifiable--it was falsifiable. The results per the protocol assessments proved that the model of the illness and treatments was wrong--or at least was not borne out in this experiment. They just disrespected their own results, as @TiredSam notes, and published bogus results.

 

Does anyone have a direct e-mail for Godwin?

 

He is a member here you could DM him ?

 

or you could pm him from here as he is a member @Mike Godwin

 

I don’t think you can pm on Twitter unless the person follows you. Ask Omar. He knows Godwin from school and has his email address.

 

Hm, is it possible to falsify a [edit: theoretical] model which is inconsistent and contradictory in itself [edit: by running a trial based on this already flawed model]?

 

Is he maybe referring to the rationale behind CBT/GET? One cannot disprove the false illness beliefs hypothesis when the patient has an unidentified illness, because disproving could only be done by identifying the actual illness and showing that the beliefs are consistent with it. Or maybe not entirely because a believer could still argue that the beliefs are driving the illness.

The CPET literature does a good job undermining the false illness beliefs hypothesis though. We can't yet identify a specific illness but the observations are consistent with patient reported symptoms and difficult to explain through false beliefs.

What would be the best way to disprove this?

 

I remember talking to Tom Kindlon way back before the PACE results came out, and it gradually dawning upon me that this £5 million trial was designed in such a way that it would be very unlikely to produce results that would change anyone's minds about the efficacy of CBT/GET. It turned out the results were even poorer than I expected, and actually do seem to have helped undermine some people view that CBT/GET are effective treatments [once the PACE spin was picked apart], but the design of PACE really didn't seem suited for moving the debate forward in the way that it should have.