I don't think that point about overlapping arms is right: according to your ref (and others I've seen) arms can overlap by half a length and still be statistically significant at 95%.Maybe you could calculate the SMC. I calculated the difference in 2F would be less than 5.5 (4mm) and in 2G around 6.9.
Somebody pointed out the following to me:
When 95% CI arms cross that can mean the p-value calculated for the difference between means will be non-significant. For example, Fig. 2B shows no overlap of CI arms at 52 weeks if you compare SMC to GET or CBT so we can conclude there is a significant difference. However, Fig 2F shows clearly that SMC upper CI arms overlap with those of GET and CBT lower CI arms. A similar pattern is also seen in Fig 2C and Fig 2G. Taking these four Figs into account, it decreased fatigue is statistically significant but there might be no significant difference in physical functioning measured.
If one looks at 2F and 2G, one seems to have enough information to say SMC isn't different from CBT and GET.
If SMC was slightly higher in 2F and 2G to 2E, it could be sufficient to make the differences no longer significant.
One caveat is that these are unadjusted data but that may not matter. I do believe their overall single p-values* look at whether individual differences are still significant at each point.
*there are four time-points and four measures at each point, resulting in lots of comparisons that might be different
As I was trying to say in the earlier post, since SMC has a higher baseline than GET/CBT you can't directly compare the 52-week end points.
IC SMC data: 52.1-39.1=+13.1
Full cohort: 50.8-39.3=+11.6 (nb lower increase for SMC full cohort)
Having done a quite a bit of work to get this data, I wish the results had been more 'interesting', but this is the way they turned out.
Intersting that the effect with fatigue appears to be quite a bit stronger than for physical functioning, which is seen in the full cohort too.
Not sure how much difference adjusting makes in practice. The difference between CBT/GET and SMC I calculated from the graph data almost exactly matched the adjusted difference figure in table 3.One caveat is that these are unadjusted data but that may not matter