The latest GET study by Prof Peter White et al is entitled Graded Exercise Therapy guided Self-help Treatment (GETSET) for patients with chronic fatigue Syndrome: a randomised controlled trial in secondary care. This trial apparently aimed to assess the efficacy and safety of Guided Exercise Self-help (GES). At the time of writing, an abstract only is available and this has been discussed on the Phoenix Rising forum. Russell Fleming summarised some of the discussion in a useful series of tweets and I asked his permission to share them here, not least because tweets are such ephemeral things and can be hard to access for future reference.
Russell frequently tweets extremely useful summaries of new research papers and notable ME-related events and is well worth following if you don’t already do so. Follow him @Firestormmer
Here are Russell’s notes on GETSET. (Whatever you think of the research, you have to admire the acronyms, though someone on Phoenix Rising suggested that GETSHT might be more appropriate.) Over to Russell:
From the discussion on Phoenix Rising and without anyone seeing the full paper, there already appears to be some issues with this trial. Quelle surprise you might think. Let me try to summarise the concerns:
1.The Chalder fatigue scale outcome was added after the trial started.
2. The NICE criteria were used to recruit but from selected secondary care centres – we don’t yet know which ones but can probably guess.
3. Only after the trial was an attempt made to allocate patients to other criteria and only Fukuda and Oxford – e.g. no mandatory PEM
4. The trial ran for only 12 weeks – which might be a fair representation of secondary care – but may not be long enough.
5. The improvements are ‘tiny’ e.g. A CFQ 6.3 point improvement means patients marked one question one step better than at baseline.
6. They also only list “adjusted” scores, so unadjusted scores probably show no improvement.
7. Serious adverse events were recorded in each arm (1% GET, 2% SMC) but we don’t know what their definition was from the abstract.
8. The pre-specified outcome measure (SF36) recorded only small effect size, so they switched to Chalder post-hoc for a better result.
9. The end-date for the trial was extended and more patients recruited while the trial was underway. This would affect outcome measures.
10. There’s mention in the Phoenix Rising discussion of a 1 year outcome measure (Chalder) but I can’t see the results in the abstract.
11. “It’s interesting how they extended the judgment period and used an additional primary outcome after the original end of the trial.”
12 There was no control group used in the trial. These authors do not appear to have learned anything from past endeavours.
Note: It amazes me that they have the cheek to conclude these measly effect sizes as ‘moderate’. What on earth would they be for recovery?!
– Russell Fleming