Yes, and as 'experience from a clinic', there were no controls, so any improvement could be, and probably was, partly natural improvement. Some people would have got better without any intervention.
Also, there was a very large dropout rate, at least 31%, and considerably more for some specific measures. And the dropouts were, on average, sicker at baseline than the completers. So, that means that just the removal of the sicker people would have been enough to result in an improvement in averages at followup.
And, of course, the measures were subjective ('patients self-reported symptomology') and therefore highly vulnerable to bias, especially given the therapy aims to convince people that they have false illness beliefs and that they can safely ignore their symptoms.