The trouble with the PACE study isn't just its flaws, but also how the researchers responded to anyone that challenged it.

For example, in order to get meaningful results from a trial like this you need to choose your definition of success before the trial starts. The plans for the PACE study did have such a definition, and while it wasn't great it also wasn't terrible. The final study didn't use it; it used a new and much weaker definition. So far this is all sadly pretty commonplace.

Then the patients involved in the study got wind of this and tried to find out what the results were if they used the original, planned methodology. According to the researchers, this was a dangerous and vexatious attack on science by activists - and because they were respectable scientific figures working for major institutions, and the people questioning their research were mostly disabled women who were, according to the mainstream consensus, nuts, people believed them.

This article is basically a direct extension of this tactic, and it's working. I can already see a few folks elsewhere in the comments calling on universities to sack the few scientists who questioned this...

What about the empirical evidence of thousands of patients who have had success with these therapies? Is it possible their experience shows it's not as simple as declaring PACE completely proves or disproves the psycho/somatic vector?

There's thought to be a natural rate of recovery of around 5%, so there will be many people who recovered while co-incidentally doing all manner of things, CBT, GET, homeopathy, etc. There's no good evidence that CBT or GET help people recover. Even relying on just subjective self-report questionnaires prone to problems with bias, the PACE trial's prespecified recovery criteria showed that there was no significant improvement in recovery rates when CBT and GET were added to patients' medical care: https://www.tandfonline.com/doi/abs/10.1080/21641846.2017.12...