This is a really trivial thing that's just been bugging me. This is how they measured fitness: results: ____________________________APT_____________CBT___________GET___________SMC only___ Edit - @Dolphin pointed out what a poor choice I'd made with the results presented, and more complete ones are available here: Few patients reported being actively harmed by the form of GET tested as part of the PACE trial, and adherence to therapy was reported as being good. Assuming GET did no real good, and any improvement in patient questionnaire scores was just down to problems with bias, I'd still expect some improvement in fitness for GET just from patients substituting in exercise and dropping off other forms of activity. I find that if I choose to prioritise exercises I can increase them by putting less energy into other tasks. Also, I often find that this does mean that I can improve my fitness and the amount I can do of a particular exercise. (This has led to me wasting much too much of my time on different forms of exercise/activity as a treatment, as for the first six weeks or so I am able to steadily increase what I can do, just from having practised a particular routine and developed particular muscles, rather than receiving any real improvements in energy levels). If GET were not actively harming patient's health (and it seems that in PACE it largely was not, even if it was wasting there time and part of a system which causes problems elsewhere), how could adherence to a programme encouraging the prioritisation of exercise not lead to improvements in fitness in a group which starts with low levels of activity? If walking was the most commonly chosen activity (I think it's been reported that it was), surely that should lead to significant benefits on the step test? Maybe I'm unusually able (for a CFS patient) to manage some sort of exercise, and these patients weren't even able to substitute activities in a way which would let them exercise for fitness while cutting out things which were actually more enjoyable/important? Maybe the therapists were being so cautious while they knew that harms were being assessed as part of PACE, that adherence to their programmes didn't really promote any additional exercise? Reference 11 was: 11. James D. Formula for fitness. University of Gloucestershire: Research Faculty of Sport, Health & Social Care 2012. So this was a post-hoc analysis... surely they could have come up with some way of claiming a significant improvement! Almost disappointed in them. [Edit - this was when I thought GET did numerically best, based on 12 weeks data - at 52 weeks GET did numerically worst] Also, I can't find this paper, or reference to it, anywhere. It must be this David James: http://insight.glos.ac.uk/academicschools/dse/staff/pages/drdavidjames.aspx It looks like this book might include some relevant info, but you cannot read much of it: https://books.google.co.uk/books?id=j3x9AwAAQBAJ&pg=PA74&lpg=PA74&dq="David James" "9.81"&source=bl&ots=lUMJH7PDL8&sig=OoNMGnI3HKjfkr6KOWa9L4r9cAc&hl=en&sa=X&ei=zZLRVPuAO4u67gbplICAAQ&ved=0CDcQ6AEwBA#v=onepage&q="9.81"&f=false Slightly regret having spent so much time on this and searching through David James' work, but just thought I'd throw it out there in case anyone else can think of something I've missed.