Bias, even if it existed in this case, is irrelevent to a certain extent.
If you find an irrefutable technical flaw in a study, it is a flaw regardless of the motivation for doing so.
Also if PACE found CBT+GET to be harmful, but significant numbers of patients were claiming to have recovered after treatment with CBT+GET and the reports of harm on patient surveys were minimal, then I can say with absolute certainty that a similar level of scrutiny would be applied as patients would be desperate for the NHS to support meaningful treatments.