UPDATE I've now seen the full text and the increase is slightly larger than I'd thought.
The findings with the aggregated scle findings (PF, pain and vitality) are probably too complicated to bother with so I'll focus on the SF36 Physical Function scores which were helpfully included for comparison with other trials (eg PACE). However, these are the 'norm-based' scores where 50= population mean and each 10 points from the mean = 1SD.
What's a little odd is that the 'control' group SF36 PF score declined slightly over the 16 months of the study. So just looking at the intervention group, from baselines:
Baseline= 32.5
16 months= 38.7
which gives a mean increase of 6.2points = 0.62 SD ie over the 0.5SD often used for 'minimally useful difference'.The increase relative to control was larger at 8.8 points, 0.88 SD.
For comparison, PACE showed a 1.2SD gain from baseline, but 0.47SD gain relative to the control group, which improved a lot. Of course, all these gains are self-reported and in both cases the control group knew they were in the control group and not receiving the intervention being tested in the trial.