Search results

Yes, the problem with the CFQ is how the questions are framed, it is a short term "how are you feeling" measure, designed to be as easily biased as possible.

Yes, as I said, he was the director of the clinic (and Clinical Trials Unit), but didn't do the database design/management himself. The bit you left out of the quote is the person responsible for randomisation, database design and management (Dr Tony Johnson).

He ran one of the clinics where the interventions were carried out, he wasn't directly involved with the design or organisation of the trial, though I'm sure he was asked for advice...

No, "reduced reassurance" relates to the reduction of a specific behaviour, namely seeking reassurance during situations that generate anxiety. In the context of health psychology, it would be about reduced seeking of medical opinions related to health anxiety.

"reassurance seeking behaviour" Example of use of "reduced reassurance" (slide 68): http://depts.washington.edu/hcsats/FCAP/resources/Modular%20CBT%20for%20Anxiety%20Disorders.pptx

Questioning the wisdom of her PhD supervisor shows genuine courage unlike another mostly well-reasoned commenter of Bad Science...

It makes you wonder though, in principle, if they treat different things, the effect should stack. We know it doesn't as the crossover shows no effect (crossover was uncontrolled in PACE trial, but still no effect).

I'd imagine that whenever the topic of the quality of the PACE trial comes up, Ben Goldacre mysteriously has to excuse himself to go to the bathroom.

SSRIs are recommended in spite of the lack of evidence from randomised control trials for any symptom, including depression as I said. It is merely (falsely) assumed that because it works for people without CFS, then it will work for people with CFS.

The analysis plan was suddenly changed when the FINE trial revealed null results. This is not a coincidence.

Basically, the establishment supporting itself. It's hard to conduct a review if you don't actually understand or engage with the specific issues at stake...

One of the problems is selection bias. Participation in studies like this tends to be skewed towards people with above average (pre-illness) IQs and hence claims about normality or lackthereof is questionable. The most valid methodology would require prospective participants before they become...

The problem with machine learning is that it isn't very useful with limited range of useful data. As Trish said, GIGO... Studies that rely primarily on questionnaires are junk, period. It's clear they want to make generalisations without any evidence...

An interesting hypothesis would investigate whether PEM related cognitive deficits are associated with short term alterations in blood volume and/or blood flow in the brain...

This is true, but it's also irrelevant when they can't objectively demonstrate efficacy for people with non-specific CF, let alone CFS or ME.

Looking forward to Gategategategate being published!

The large unexplained variance is due to the unreliability of the questionnaire itself. This is not specific to the questionniare used (it may be one of the better ones), but an inherent flaw of questionnaires in general.

How about a third investigative reporter reporting on the second? We can call it Gategategate.

I found this curious. The risk of autoimmune diseases in first degree relatives is discussed here: https://genetics.emory.edu/documents/resources/Emory_Human_Genetics_Autoimmune_Disorders.pdf Two out of my three siblings have autoimmune diseases (Type 1 Diabetes, autoimmune Thrombocytopenic...

I agree, it is a perversion of the term. A Randomised control trial requires a (patient) blinded control arm.