Thanks a lot for doing that. I thought there were a lot of good points included.
Some bits I wasn't so sure about, eg:
One of the biggest flaws in the study design was that it was not blinded, meaning both the participants and the providers knew which treatment they were receiving, and this could have led to an over-estimate of outcome effectiveness.
I understood what you were getting at here, but thought that the phrasing made it sound less persuasive than it could have been. As Jonathan Edwards makes clear, it's the combination of a nonblinded trial and subjective self-report outcomes that are the real problem. The nature of an intervention like LP means that problems with bias really cannot be avoided for subjective self-report outcomes. The document does follow up by talking about potential use of a sham therapy control group, and that's a sensible point, but I thought the above sentence could have been better phrased.
It appears to be a form of intensive CBT (cognitive behavioural therapy), used to correct what are regarded as negative thinking patterns, teach stress-reducing techniques, and to set – and encourage people towards – personal goals of improvement.
But it remains unclear the extent to which NLP, osteopathy and hypnotherapy come into it, although anecdotal reports provide some worrying clues.
That description of LP makes it sound rather more respectable than it is imo (although I realise that it's hard to get clear information on exactly what the LP is).
Three of the participants allocated to the SMC-only arm went on to seek private LP.
This may confirm that they only entered the study in the hopes of getting LP for free or could reflect how well LP was portrayed in the initial information given prerandomisation.
It is not clear whether the data from these 3 participants was then used in the SMC or SMC+LP data at the end, but it could have affected the results.
I think that they were included in the SMC only group, which I think is the right way of doing it for their primary analysis - sticking to their allocation per randomisation. There was some discussion of this in the thread on SMILE. I think that this is also relevant to the third additional concern mentioned. I think that these numbers on this do add up okay, although it would be nice to have more details included.
This meant that the participants could ‘pick and mix’ the parts they felt they wanted to receive and so it was not consistent across all participants, making for a very un-controlled control group!
I'd have liked more information on what was actually provided as a part of SMC, and particularly if there ended up being any group differences in what was provided, but I also think it's okay to have a group where participants are able to choose what aspects of the available treatment they made use of. If 'standardised medical care' tends to allow this sort of personalisation in routine clinical practice, why shouldn't it be able to in a trial (so long what is involved is properly reported).
This trial determined that the Lightning Process® was effective when used in addition to specialist medical care in the treatment of ME/CFS in children.
Did it? I can see Phil Parker being pleased with that endorsement from the MEA.
As I feel bad only pulling out the bits I had concerns about, I thought I'd also highlight this good summary which was near the end of the document:
Children may be more impressionable than adults and may have a stronger desire to please their therapist (or their parents) or to answer questions in a way which reflects what they have been taught e.g. to misrepresent their symptoms, which was a particular concern with the Lightning Process®.
In addition, it has been shown that children are more likely to make a full recovery and recover more quickly than adults (Jordan et al. 2010, Burgess 2011, Norris et al. 2017).
The children in this study could have improved naturally, which may have contributed to their improved outcomes, however, any such improvement has been assumed to be the result of an intervention.
A proper control group – a non-treatment group – would have helped us to see if children had improved without intervention and we could have made appropriate comparisons.
