I think we used "model" in the paper.
".. interventions were built upon a behavioural/deconditioning model of CFS."
In psychology, to avoid this kind of confusion, we tend to only use "hypotheses" for the fully operationalised version(s) of the prediction(s), the ones that are actually put to the test statistically.
I think they refer to a model in their CBT and GET manuals. But I don't think I would call it a model as its a few boxes with lines between them. That is I don't think there is enough to tie down what they are saying.
I agree that "model" is probably a misnomer, and as @Robert 1973 points out, a lot of the slipperiness is around the imprecise use of language. But I would think just "model" or "framework" as a shorthand term in informal public discourse is nonetheless preferable to "theory" or "theoretical framework". Inclusion of a modifier clarifying substance and precisely synonymous with "hypothetical" would be preferable, but I'm afraid I'm unable to offer a decent candidate.