Coming back to epistemology. My conclusion from this is that the story with ME/CFS is not a balance of two competing theories of causation, as you suggest in your lecture, but a huge imbalance between
OPINION
1. Wessely school opinion about the efficacy of CBT and GET as treatments.
JUSTIFIED BELIEF
2. PACE, FINE and other trial evidence that CBT and GET don't work and patients' experienced evidence of harm from these treatments.
I think the problem here is that you appear to be approaching a problem philosophically when it's already been solved scientifically. Research by proponents of CBT and GET demonstrates that those treatments result in no objective improvement when objective data is collected and reported. Those CBT/GET proponents deal with that evidence by explaining that the objective outcomes don't matter - only self-reported fatigue is relevant.Okay, but that distinction to me is the ballgame, and I lack the relevant expertise to safely rule that one is just opinion, and the other a justified belief.
And even the subjective improvements are questionable, due to poor design of CBT/GET trials, and the heavy emphasis on encouraging patients to treat their symptoms as being irrelevant. For example, in the Netherlands, the official CBT/GET guidelines prominently feature that the patients should be taught not to view themselves as patients. Then trials incorporating those guidelines will use the answer to the question "Are you still a CFS patient?" (or similar) to declare that CBT/GET has cured the patient. Hence the very process of CBT/GET revolves around inducing bias, which is specifically exploited by some of the questionnaires used to determine improvement, and is likely responsible for most or all of the positive effects seen after treatment.
The substantial nature of the role of bias in CBT/GET trials is also likely illustrated by the long-term followups, when conducted and reported. At that point, there is no longer any apparent benefit to having had CBT/GET, and even the most pathetic of control arms have caught up. Even if psychosocial researchers were correct in asserting that only fatigue matters, and not functionality, the effect of CBT/GET on self-reported fatigue wears off after a year or so, which suggests that the short-term fatigue improvements were entirely a result of bias, and the objective outcomes are more accurate and more predictive of long-term effects.
Harms from CBT/GET are indeed trickier, because the CBT/GET proponents do a piss-poor job of collecting such data. And when there is any collection of harms data, it's done in a manner to minimize the reporting of post-exertional malaise - for example, by declaring that an adverse event must last at least two months to be "serious", and by grouping together everything else as non-serious. Thus being bed-bound for a month following a week of GET would be buried in a mound of hundreds of minor and incidental symptoms. And even in the PACE trial where that burying of PEM happened, GET had significantly more Serious Adverse Events - until therapists unblinded the patients and decided that the adverse events couldn't be related to the therapies.
So with a lack of good evidence from trials, which we would all prefer, we fall back on huge patient surveys. A fairly large majority of patients report being harmed by GET, and very few report finding any benefit from CBT and GET. We can also look at the studies from teams of exercise physiologists around the world which show that ME patients have a decreased VO2max 24 hours after a prior maximal CPET, something which happens in no other known group, including sedentary, deconditioned, or ill controls who don't have ME. While not conclusive proof that exercise is harming us, it does suggest that patients are correct in asserting that exercise makes them very sick in a way which doesn't even remotely resemble deconditioning. At the very least, that should shift the burden onto the GET proponents to carefully and thoroughly prove that exercise is safe despite those demonstrated physiological abnormalities.