A good question, and one that the NHS community is aware of and trying to address.
We don't know for sure, but I suspect a few things are at play:
I suspect therefore that in some cases, what is being provided by the NHS is different from the PACE trial, and therefore can't be easily compared. Some NHS services are not able to provide the 15 one-to-one (one hour) sessions as undertaken in the PACE trial, for example. The PACE trial also had a very high level of training and supervision for the research therapists, which is not always easy to replicate in the NHS. It's also sometimes not straightforward to distinguish between GET/CBT and APT in clinical practice, whereas this was very carefully separated during the PACE trial. This is my personal hypothesis, but would need further exploration.
One thing we are doing via BACME is to put together training for NHS therapists nationally that more closely replicates the PACE protocols than training undertaken previously: this ensures that the new learning from the PACE trial is disseminated more widely.
BSc (Hons) Physiotherapy, MCSP, PVRA, HG (Dip), MBACME