Efficacy of web-based cognitive–behavioural therapy for chronic fatigue syndrome: randomised controlled trial (2018) Knoop

thank you so much for that laugh @Jonathan Edwards & @Valentijn . I was just feeling so demoralised & infuriated, feeling that this tide of bullshit will never end, i needed that giggle so much

Dont forget, that PACE was "a great great trial"

 

They combine CBT with a GET component. So they say CBT is effective but really its CBT and GET.

 

Thanks, where exactly does it say they use GET also?

 

I think it suggests they daren't try the same thing with GET.

 

Can these two threads be merged?

 

Definitely a trial ...

 

That's a good question. I haven't read this paper but normally Blijenberg will do stuff like that. I've seen it in other studies like this one. Although it's not for CFS.

https://www.researchgate.net/public...uscles_of_active_FSHD_patientsAuthor_Response

 

It may take less time because, as they are saying the same thing to each victim, it can be a video, so one video, say an hour, for as many victims as they can scam, per session, as opposed to the old model where the time needed was 1 hour per victim per session.

So not less time for the victims, just much, much, much less for the "therapist", and thus cheaper.

 

Even if we assume that the statistics are correct, I have questions.

So patients had to:

• be well enough to make (and keep) at least one appointment outside the home.
• able to speak, read, and write
  
• be able to do the above in Dutch
• able to use a computer
• have access to the internet
• be able to write e-mails

So the selection criteria weeds out

• the immobile,
  
• the cognitively challenged,
  
• the computer illiterate,
  
• people too poor to own a computer
  
• or have access to the internet,
  
• people who cannot use computer programs
• people who couldn't manage a non-user generated username or password?
  
• people who don't speak Dutch as a first language or struggle with it
• People involved in legal disability claims (people who are willing to challenge authority?)

So, the criteria for being involved probably weeds out people who are probably the sickest and poorest? (and possibly most likely and able to question authority?).

So you didn't check:

• If the module was read, only opened?
  
• Why?
  
• No mechanism at the end to show completion of the module?
  
• Why?
  
• How do we know the patient actually received the treatment then?

Why 5%? No evidence shown for how this was decided or validated.

Again, why? You cannot measure internet activity by recording the number of times patients logged in, mean duration of sessions, number of opened treatment modules, and number of emails sent.
Assuming UX/UI (User eXperience/User Interaction) assumptions are true, i.e that web companies do not spend a lot of money and research on UX/UI for no reason, that the technology for measuring UX/UI are not useless, etc.

• My current computer session is 8 days, only an idiot would infer that I've been actively using constantly for 8 days without a rest.
  
• What were the settings for session time-outs? Where there any?
• Why do these patients not exhibit the web using trends of general users?

I'm only at the end of pg2 with the questions! I'd be interested to hear more from just the technology side.

How do all of these factors constitute a representative(?) sample?

Questions for therapists: How important is non-verbal communication (and/or non-speech language) in assessing a patient? Why isn't it relevant in iCBT?

 

Additional questions:

• where is the 'web portal'?
• can it be audited?
• what accessibility standards were used in the design?
• who approved them?
  
• can we see the raw (anonymised) usage logs?

 

The disability claims exclusion criteria is odd, and that seems worth querying, but given the nature of the intervention there were always going to be limitations on who can take part. I don't think that most of them are a real problem with the trial. In a trial of iCBT, providing a computer and computer training to those who would need them is quite an extra task (and one unlikely to be taken up in clinical care outside the trial).

 

How do we even know that the patients received some or all of the treatment? We know that they started 'treatment modules', but there is no proof that they completed them. What if they felt an obligation to say they completed modules because they wanted to please or look like they tried?

Technically we don't even know the contents of the e-mails, just that they exist?

 

A wait list control is pretty poor. I read this paper really only interested in their actimeter results, but not much info was given on them. I think that they may have messed up their supplementary files. They also mentioned doing a range of analyses, but I couldn't see the results for a lot of them. It is past my bed-time though, so I may have just missed them in my rush.

I pulled out some bits that I thought were of interest, but this was only a quick read through:

I feel like they should have explained that this related to depression and anxiety, not CFS.

This was part of their entry criteria:

The average SF36-PF score for all three groups was 62 at baseline. ie the average participant fulfilled the PACE trials SF36-PF recovery criteria (although this is also under the PACE trial's entry criteria for severe and disabling fatigue, which required and SF36-PF of 65 or under). Presumably many of the participants had scores significantly higher than this too.

I've not looked at the reference for this, but it's from 1988 and it doesn't mention CFS in the preview page that was available:

Interesting that they gave no justification for this cut off:

They mentioned these changes from their protocol (protocol is here, I've not read it: https://bmcneurol.biomedcentral.com/articles/10.1186/s12883-015-0392-3 ):

I thought it was a bit fun that they took such care with this, but then had a wait-list control:

I wonder why this was?

Maybe this is where the 40% recovery claim came from? I wonder if there was a press release for this paper:

Yeah - this was from later on:

They said:

This is mentioned here, but I can't see supplementary Table 2 in the supplementary files.

I don't really understand what was happening with their actigraphy, why they were not planning to assess results, and why they then did. The mentions it gets:

I can't see supplementary Table 1 either. I've got table three, but it looks like they may not have uploaded the final file:




I think that this is as close as they come to addressing criticisms (ref 32 is PACE recovery, 34 is PACE Lancet):

This is what they say on physical functioning:

I let Unrest run on Netflix while I was doing something else, just to indicate an interest. It's hard to check on some of this stuff. They gave some more details on this here:

 

Oh - and in their third supplementary material, they did describe their CBT as having graded activity - I saw discussion of that somewhere:

 

It's important to filter out people who think and fight for themselves - which is one mission of psychiatry. Furthermore, people who are involved in legal disability cases probably know they have a physical illness, and that CBT is useless. It's harder to convince them of the efficacy of CBT.

To me, you rise good points, and they make this publication a low-quality study in my opinion. On top of subjective outcomes you have these issues.

 

This seems to be saying that full adherence was very low?

 

Oh, seems so. 16% and 19%, yes? Are there reasons given?

What is "protocol-driven" and "on demand"? Does protocol-driven mean: Click according to a plan; and on-demand: whenever you feel like it? - Ah, yes.