How to design a good clinical trial of brain retraining

UkPoster said:
Utsikt, I'm so curious about how you would design a brain retraining experiment. In a way that you would think was a fair test of it. Because my understanding is it's hard to do because you need to do it in a way the kindness of the therapist and their attention isn't a factor. That's what others have said to me anyway, just doing a treatment trial wouldn't show anything.
Click to expand...
One of the simplest things to do is to use objective outcome measures, in particular for something like CBT as designed in ME/CFS trials. For example long-term activity monitoring. That captures only one facet, but an important one that contradicts the claims made by people offering CBT: Do people become more active. We know that for CBT that answer is: No.

Regarding trial setup to counter kindness and allegiance, I would think the simplest thing to control for therapist and investigator allegiance to me would be to have a control group, where a drug that is believed to have an effect by the practitioner, take LDA for example, is handed out by said practitioner and use that as control. That would allow you to control for those factors. However, additional factors exist. Randomization and non-blinding leads to some bias but I think that's where a crossover could be useful. I think the larger problem would then be the disparity in attention and time received in the different treatment arm(s). I would think this can be countered by introducing something like supportive counselling in your control arm(s). Supposed drug effects can obviously be handled in subsequent placebo-controlled trials. However, using objective outcome measures is already more than what people running such trials are typically able to handle.
 
EndME said:
One of the simplest things to do is to use objective outcome measures, in particular for something like CBT as designed in ME/CFS trials. For example long-term activity monitoring. That captures only one facet, but an important one that contradicts the claims made by people offering CBT: Do people become more active. We know that for CBT that answer is: No.

Regarding trial setup to counter kindness and allegiance, I would think the simplest thing to control for therapist and investigator allegiance to me would be to have a control group, where a drug that is believed to have an effect by the practitioner, take LDA for example, is handed out by said practitioner and use that as control. That would allow you to control for those factors. However, additional factors exist. Randomization and non-blinding leads to some bias but I think that's where a crossover could be useful. I think the larger problem would then be the disparity in attention and time received in the different treatment arm(s). I would think this can be countered by introducing something like supportive counselling in your control arm(s). Supposed drug effects can obviously be handled in subsequent placebo-controlled trials. However, using objective outcome measures is already more than what people running such trials are typically able to handle.
Click to expand...

@UkPoster all of EndME's suggestions are really important and as others have said using objective outcome measures is necessary because you can't blind. If any planned trials are unblinded with subjective outcome measures then it is guaranteed to be unable to tell us anything before it even starts. Other possibilities for objective outcomes could be eg 'days spent in work.'

I would also suggest it's important that the two groups are matched in terms of 'duration of illness.' We have hard data from Kielland's work on Norwegian population data that recovery from ME/CFS almost never happens after a few years of being sick. However natural convalescence seems likely to be really high in those who have only been sick a handful of months - though we could do with better epidemiology on this.

@UkPoster you made a comment somewhere before about claiming this community is like anti-vaxxers which I think is really unhelpful. There is overwhelming evidence of the efficacy of vaccines for many diseases - whereas psychological and behavioural interventions have repeatedly failed to show any objective improvements in ME/CFS. Everyone here has sincerely tried all sorts of permutations of CBT, lightning process etc (I have personally done both) and only end up here after they have done nothing at all for their illness or encouraged them to do activity which has made them worse.
 
A belief can give us the strength to pursue an idea. But if the idea has no merit, then the belief is harmful.

We also need the competence to determine whether an idea has merit or not.

Competence makes the difference between a healthy, useful belief and an unhealthy one that leads us into failure.

I don't see much competence in treatments like the Lightning Process. It looks just like cynical exploitation of patients: critical thinking is discouraged while strong beliefs are encouraged. The proponents who claim to believe in their treatments avoid conducting clinical trials that could tell us whether they really work. The underlying theories look like popular neuroscience and psychology. Allegedly the unproven treatments work for a wide range of poorly understood conditions. Proven treatments tend to work only for specific problems, there are no universal cures.
 
Last edited:
UkPoster said:
Utsikt, I'm so curious about how you would design a brain retraining experiment. In a way that you would think was a fair test of it. Because my understanding is it's hard to do because you need to do it in a way the kindness of the therapist and their attention isn't a factor. That's what others have said to me anyway, just doing a treatment trial wouldn't show anything.
Click to expand...
The others have already provided some good suggestions.

The first step would be to build a coherent theoretical basis in a way that provides testable hypotheses that can only be tested with a trial. As demonstrated in this thread (edit: see the thread quoted in the first post of this new thread), PP (edit: predictive processing) and BR (edit: brain retraining) is nowhere near that stage yet, because there are so many contradictions.

Once you have a hypothesis (or a few), you need to design a trial that controls as many variables as possible. Like you mention, attention and kindness is one of them. Another, and maybe the most difficult to control for BR, is that BR attempts to directly manipulate many of the variables you might want to measure. It tells you how to think, speak and behave. This means that a change in those variables might represent adherence to the intervention, and not an effect on the disease.

One way you might try to get around this is to use long baseline and followup measurements for objective outcomes, like what Fluge and Mella are doing in ResetME.

If you hypothesis is that BR will lead to a sustained increase in physical activity without a decrease in quality of life or ADLs you can do, and without an increase in symptoms or other complaints, then you need to demonstrate that over e.g. 2 years first to show people are not just pushing through, and preferably with 5+ years followup like what you do for e.g. cancer treatments or treatments for progressive diseases.

A step counter can be used as a proxy for physical activity, and you can track income and welfare benefits, employment status and hours worked (or studied), and medications. Objective measurements of mental and physical ability can also be used, assuming they are PEM-safe.

To not rely on the patient only, you can ask for assessments by their family members as well.

The largest challenge for BR specifically, is actually that a proper trial requires that you fully acknowledge that the treatment 1) is experimental and based on unvalidated theories, 2) has substantial reported risks, 3) might not work, and 4) that it’s not the patients fault if they do not improve.

The control group can’t be pacing, because pacing isn’t a treatment so it won’t generate the same expectation of benefit, and comparing it to something that’s not PEM-safe won’t be ethical. (Edit: this assumes that BR would be PEM-safe, which is by itself a big assumption, possibly making any BR trial unethical by nature.)

One alternative is to do as Wyller did in MINIRICO. They compared the supplement NR with MBRT, their version of LP and BR. They used four groups:
a) MBRT and NR;
b) usual care and NR;
c) MBRT and placebo;
d) usual care and placebo.

According to their recent grant application, NR as ineffective and MBRT failed to reach the threshold of minimally clinically important difference for the subjective outcomes and had no change for objective outcomes. That did not stop them from claiming success and securing a £2M grant for yet another study of MBRT. Meaning that funding is easy to get for these kinds of trials. Resources is not an issue here.

We know from Ritux that you can generate considerable bias with an ineffective drug for ME/CFS, and the recent LDN trial presented at Charité shows the same for LC. That would make a similar treatment a potentially viable active control.

Recruitment bias might be an issue, but it could be somewhat alleviated by the possibility of getting the exciting drug if you’re not into BR and wouldn’t have participated in a pure BPS trial.

You would do best by recruiting a fairly homogenous group, and avoiding ceiling or floor effects by excluding the most and least healthy.

I’m sure there are other things to consider as well. My approach to trying to find them is to ask myself how I would try to poke holes in a study if it used the same design and had a result I really did not like. If a biased me can’t find any flaws, a neutral me shouldn’t either. And of course ask others for help doing the same - just make sure they are not biased in favour of you or the intervention(s).
 
Last edited:
chillier said:
@UkPoster all of EndME's suggestions are really important and as others have said using objective outcome measures is necessary because you can't blind. If any planned trials are unblinded with subjective outcome measures then it is guaranteed to be unable to tell us anything before it even starts. Other possibilities for objective outcomes could be eg 'days spent in work.'

I would also suggest it's important that the two groups are matched in terms of 'duration of illness.' We have hard data from Kielland's work on Norwegian population data that recovery from ME/CFS almost never happens after a few years of being sick. However natural convalescence seems likely to be really high in those who have only been sick a handful of months - though we could do with better epidemiology on this.

@UkPoster you made a comment somewhere before about claiming this community is like anti-vaxxers which I think is really unhelpful. There is overwhelming evidence of the efficacy of vaccines for many diseases - whereas psychological and behavioural interventions have repeatedly failed to show any objective improvements in ME/CFS. Everyone here has sincerely tried all sorts of permutations of CBT, lightning process etc (I have personally done both) and only end up here after they have done nothing at all for their illness or encouraged them to do activity which has made them worse.
Click to expand...
Really?
That’s rich because the brain training gang are the anti-vaxxers in this scenario. Rejecting orthodox science and biological mechanisms in favour of their belief in untested and unresearched non-scientific practice!
 
Last edited:
From my observations and reading 'recovery stories' has solidified my beliefs that most people recover from viral infections within a year.

The fact that I recovered from Covid around the six month period, and still have ME/CFS delayed PEM for over 30 years, tells me to stop wasting energy on these BR programs for LC and ME/CFS.
 
You must log in or register to reply here.
Back
Top Bottom