FITNET Trial: Effectiveness of internet-based CBT for CFS: an RCT (Nijhof et al, '12)

[ukxmrv]

At least you give a clue on what to look for Willow. One of the Dutch patients may be able to confirm what is considered a normal school attendance for them. Thanks for posting that.

 

[charityfundraiser]

Don't know if this has been posted before but here is a table containing CIS-20, CHQ, and school attendance of healthy controls and adolescent CFS from the paper cited in the FITNET paper regarding:

"Recovery was defined post hoc, in relation to healthy peers (2 SD), as having a fatigue severity score of less than 40, physical functioning score of 85% or more, and school absence of 10% or less in the past 2 weeks, all assessed at 6 months."

Table: Morbidity of Adolescents With CFS Reported by Pediatric Departments in 2008
http://pediatrics.aappublications.org/content/127/5/e1169/T1.expansion.html

(I made a correction here on March 6)
CIS-20 score and school attendance of those who recovered on FITNET at 6 months (2 SD) is the same as the scores of these healthy controls. CHQ score is a little lower. The FITNET paper appendix/supplementary material doesn't give the average scores for the 1 SD recovered patients, just the number of patients beyond the cutoff. I was originally looking to compare the average CIS-20, CHQ, and school attendance scores between the 1 SD recovered patients and healthy controls to see what scores the statement "I am completely recovered" corresponded to but they don't make this possible the way they presented the data. Given that the scores are the same as or worse than the scores of healthy controls for the 2 SD recovered group, then I assume the scores are slightly better for the 1 SD recovered group. However, only a few patients fell in between the 1 SD and 2 SD cutoff so this effect is probably small.

(I made a correction here on March 6)
On the other hand, about 2/3 of the patients considered recovered by the 1 SD cutoff considered themselves "completely recovered". This is even though the 1 SD recovered group scored the same as healthy controls on the CIS-20 and school attendance. I don't think 1/3 of healthy controls with the same scores would call themselves the equivalent of not completely recovered.

Or who knows.. "healthy" people really vary. One former friend I roomed with who was "healthy" had about the same activity level as I did when I was mostly bedridden. She just reclined on the couch and watched TV for 14 hours per day, napped a lot, and the most exercise she got was some really slow walking. She did however have mental problems, undiagnosed as they may have been. So sometimes I can understand why people think that's what CFS is because that kind of people and mental disorders do exist.

Anyways, 2 SD.. I think in a normal distribution, that means about 5% falls beyond the cutoff, including upper and lower ends, so assuming only the lower end applies to CFS patients recovering compared to healthy controls, that means the 2 SD recovery cutoff corresponds to the 2nd percentile of "healthy" controls, if I'm understanding this correctly, which is borderline not healthy, but I guess that actually makes the cutoff similar to the healthy/not healthy distinction.

 

[oceanblue]

However, another point to all this is that it's just one moment in time in a relapsing-remitting condition.

I guess this could be regression to a mean, where patients who wouldn't normally be severe enough to qualify for the trial are included if they happen to be in a relapse, then appear to 'improve' as they recover from the relapse independently of any treatment. The converse, counterbalancing example is a patient who would normally be severe enough to qualify, but didn't get selected because they happened to be in remission. But because they aren't selected for the trial you lose the counterbalancing effect.

This effect should apply equally to control and treatment groups in a randomised trial, at least on average, but may be the explanation for some unusually good results.

 

[oceanblue]

Don't know if this has been posted before but here is a table containing CIS-20, CHQ, and school attendance of healthy controls and adolescent CFS from the paper cited in the FITNET paper regarding:

"Recovery was defined post hoc, in relation to healthy peers (2 SD), as having a fatigue severity score of less than 40, physical functioning score of 85% or more, and school absence of 10% or less in the past 2 weeks, all assessed at 6 months."

Table: Morbidity of Adolescents With CFS Reported by Pediatric Departments in 2008
http://pediatrics.aappublications.org/content/127/5/e1169/T1.expansion.html

CIS-20 score and school attendance of those who recovered on FITNET at 6 months (2 SD) is the same as the scores of these healthy controls. CHQ score is a little lower...

On the other hand, only about half of the patients considered recovered by the 1 SD cutoff considered themselves "completely recovered". This is even though the 1 SD recovered group scored the same as healthy controls on the CIS-20 and school attendance. I don't think half of healthy controls with the same scores would call themselves the equivalent of not completely recovered.

Anyways, 2 SD.. I think in a normal distribution, that means about 5% falls beyond the cutoff, including upper and lower ends, so assuming only the lower end applies to CFS patients recovering compared to healthy controls, that means the 2 SD recovery cutoff corresponds to the 2nd percentile of "healthy" controls, if I'm understanding this correctly, which is borderline not healthy, but I guess that actually makes the cutoff similar to the healthy/not healthy distinction.

Thanks for that. I thought it might help if I put the data you used below:

	Attend	Function	Fatigue
FITNET, Recovered, 12months: mean	96.3%	89.8	21.0
FITNET, SD	8.6	18.6	10.7
'Healthy' population mean (Nijhof 2012)	98%	95	30
healthy, SD (1)	4	5	5
Healthy mean (Nijhof 2011)	98.5%	95.1	22.6
healthy , SD	3.9	7.6	10.6

nb the mean/SD from that 2011 study you quote is slightly different from the ones they use in the current paper. Here's what the 2011 paper says about a) the data used for the entry criteria (mean -2SD) and b) the control group used for their study, quoted above. They seem to be using different means/SD every time, which makes life confusing:

The cutoff point for impaired physical functioning was set at ?85 (healthy population's mean of 96.8 minus 2 SDs [2 5.4]).19 As a reference group, students at a Dutch secondary school (de Breul, Zeist) were invited to participate. Adolescents who were suffering from a chronic illness and adolescents currently under treatment were excluded [my note: ie a proper healthy population, wiht chronically ill excluded]. The participation rate was 85%. From this group of students we randomly selected individuals for a control group matched for age and gender (n = 144; mean age: 15.3 0.6 years; 79% girls).

Also, I'm not sure if the 'healthy' controls used in the current (2012) study were specifically healthy (excluding those who are chronically ill), it seemed they were merely well enough to attend school (not necessarily all the time). Using 2SD on such a population is likey to take in unhealthy individuals, making it an inappropriate threshold for recovery. PACE did something very similar with their 'normal' definition.

 

[Vitalic]

Let's be honest, if these kind of results were found in a study that was looking at bio-medical causes we would stand up and take notice (as has happened with the Rituximab trial). This seems to be far more clear cut than the PACE trial, whatever the cause of this illness is, it seems that there is some subset of people who suffer with idiopathic chronic fatigue for whom this kind of intervention is effective. I feel both sides of the argument are culpable in preventing progress, because one side is so adamant that the cause is not psychological that they flatly dismiss any finding which would implicate that hypothesis, but the other side doesn't even acknowledge that there may be a spectrum of different illnesses, and that applying the results of studies done on patients with idiopathic fatigue to people with chronic ME/CFS is misleading and utterly useless. I think the future is very bleak indeed if we can't separate and define these illnesses properly.

 

[Esther12]

Let's be honest, if these kind of results were found in a study that was looking at bio-medical causes we would stand up and take notice (as has happened with the Rituximab trial). This seems to be far more clear cut than the PACE trial, whatever the cause of this illness is, it seems that there is some subset of people who suffer with idiopathic chronic fatigue for whom this kind of intervention is effective. I feel both sides of the argument are culpable in preventing progress, because one side is so adamant that the cause is not psychological that they flatly dismiss any finding which would implicate that hypothesis, but the other side doesn't even acknowledge that there may be a spectrum of different illnesses, and that applying the results of studies done on patients with idiopathic fatigue to people with chronic ME/CFS is misleading and utterly useless. I think the future is very bleak indeed if we can't separate and define these illnesses properly.

Speaking for myself, I was pretty hesitant about getting excited over the Rituximab trial, and critical of some the presentation of results to the media... not least because I saw how the results from subjective questionnaire scores were used in PACE. I remember a number of others being the same, even with the advantage of blinding for the Rituximab study.

To some extent, I think that the history of misleading claims justifies greater scepticism of this paper, particularly with Bleijenberg's name on it.

However I also agree that it does seem as if these results indicate a really significant effect from FITNET... although there are still lots of question marks over it. This study produced much more significant results than PACE, but there are problems with it, and particularly the control group used. That the study was aimed at adolescents also means that I'm less able to interpret the data than I would be with adults (adolescents diagnosed with CFS tend to have higher rates of natural recovery... but I'm not at all well read on this).

There probably is more instinctive scepticism of this result that the rituximab one, partly due to some hostility towards those who have promoted a psychiatric approach to CFS. While I think this is fair, given the harm and problems which they have caused and the way in which papers were written misleadingly in the past, I also agree that we need to be careful to ensure that it does not end up distorting the conclusions we reach once the paper has been more carefully examined.

I completely agree about the need to work on subdividing patients, and only apply treatments to them when there is good positive evidence that they are relevant to the particular patient, or as part of research trials and with informed consent.

 

[Valentijn]

Let's be honest, if these kind of results were found in a study that was looking at bio-medical causes we would stand up and take notice (as has happened with the Rituximab trial).

We are taking notice, obviously, since we're discussing it. We've also learned to be very wary of anything coming out of Nijmegen, due to the absurd conclusions they draw from data, and their history of failing to mention data that contradicts their statements of success. There have also been past claims from at least one Nijmegen patient that his bad results in response to their therapy were discarded.

I'm not sure where they're hiding the ball in this study, but I'd be surprised if everything is exactly as it seems, based solely on their history. Close scrutiny is certainly warranted.

 

[Dolphin]

However, another point to all this is that it's just one moment in time in a relapsing-remitting condition.

I guess this could be regression to a mean, where patients who wouldn't normally be severe enough to qualify for the trial are included if they happen to be in a relapse, then appear to 'improve' as they recover from the relapse independently of any treatment. The converse, counterbalancing example is a patient who would normally be severe enough to qualify, but didn't get selected because they happened to be in remission. But because they aren't selected for the trial you lose the counterbalancing effect.

This effect should apply equally to control and treatment groups in a randomised trial, at least on average, but may be the explanation for some unusually good results.

Thanks for exploring that - regression to the mean is one point.

However, there is in my view a difference between "improvement" and "recovery". If you recover from a condition in my book, you don't have it again. It's history. Being well for a period in a relapsing-remitting condition (in my book) isn't recovery esp. if the person isn't being "stressed". If they can play competitive sports for a year, say, I'd be more happy with that being called recovery.

I was in full-time education for over 4 years with the condition. Even managed to build up to a reasonably large amount of exercise (1000m swimming or 6 miles cycling every two days). But I couldn't live the life I wanted to live. I couldn't have much of a social life. I couldn't live the life of a young person. I had some cognitive problems. I wasn't recovered. Occasionally (I wasn't undiagnosed) I'd play an (uncompetitive) game of football and I'd feel terrible the next day - I couldn't do it. Eventually I relapsed. And am now severe. By their criteria, I almost certainly would have been classed as recovered.

 

[Dolphin]

(To nobody in particular)
Saw this posted on a discussion on the CAA's FB by Danny Ze-dog:

The danger of relying on self-report rather than objective measures to assess physical function has been noted many times before. For example, Barber (1991) studied the correlation between subjective health questionnaire reports and cardiac function in adolescent patients and found that "questionnaire data significantly overestimated exercise ability in 67% and underestimated it in 3% of the subjects".

He gave more info:

Barber G, Heise CT: Subjective estimates of exercise ability: Comparison to objective measurements. Ped Exerc Sci 3:327-332, 1991

Another one you may find interesting, which used the same questionnaire and found the results did not correlate with cardiac function:

Ginsberg JP, Cnaan A, Zhao H et al. Using health-related quality of life measures to predict cardiac function in survivors exposed to anthracyclines. J Clin Oncol 2004;22:3149-3155."

 

[charityfundraiser]

oceanblue - Thanks for doing the table again. Could you check the CHQ scores for the recovered and controls in your table? I think they are reversed, or my eyes are crossed. Anyway, I guess the school attendance was actually slightly higher in controls and the CIS-20 score was actually slightly higher (worse) as well, but seem close enough that I called them the same earlier.

And I need to correct my earlier statement that "only about half of the patients considered recovered by the 1 SD cutoff considered themselves "completely recovered". It is actually about 2/3. What I was looking at earlier for the denominator was the number of patients who fulfilled at least one of the score criteria (school attendance, CIS-20, or CHQ) instead of all three.

Oh and I need to correct this as well. My logic yesterday was the reverse of what it should have been.

"Given that the scores are the same as or worse than the scores of healthy controls for the 2 SD recovered group, then I assume the scores are slightly better for the 1 SD recovered group. However, only a few patients fell in between the 1 SD and 2 SD cutoff so this effect is probably small."

 

[Adster]

Sorry if this has been posted before, or if this is not the right place. This Australian radio program appears to be discussing this study http://www.abc.net.au/radionational/programs/lifematters/web-based-therapy-for-cfs/3865870

 

[Dolphin]

Thanks, Sam.

Subjective Estimates of Exercise Ability: Comparison to Objective Measurements


However, a couple of more recent papers on adolescents and adults with heart disease found decent correlation of r = 0.521 and r=0.435 (see fig 3), with Physical Functioning questionnaire scores. I'm also not sure how precise a correlation there should be between general physical capability (e.g. ability to climb a flight of stairs) with peak oxygen uptake.

but it's certainly useful evidence of the problems of relying on questionnaires to assess physical function.

I'm willing to believe activity questionnaires can have a value in people who don't have ME/CFS and who haven't been put through particular programs.

It's quite different I think where there is a real possibility of a ceiling of activity in ME/CFS and so where there is a real possibility that substitution may be taking place with some programs.

 

[oceanblue]

However, there is in my view a difference between "improvement" and "recovery". If you recover from a condition in my book, you don't have it again. It's history. Being well for a period in a relapsing-remitting condition (in my book) isn't recovery esp. if the person isn't being "stressed". If they can play competitive sports for a year, say, I'd be more happy with that being called recovery.

I was in full-time education for over 4 years with the condition. Even managed to build up to a reasonably large amount of exercise (1000m swimming or 6 miles cycling every two days). But I couldn't live the life I wanted to live. I couldn't have much of a social life. I couldn't live the life of a young person. I had some cognitive problems. I wasn't recovered. Occasionally (I wasn't undiagnosed) I'd play an (uncompetitive) game of football and I'd feel terrible the next day - I couldn't do it. Eventually I relapsed. And am now severe. By their criteria, I almost certainly would have been classed as recovered.

I certainly agree there is an important difference between recovery and improvement. I take you point that although you could do quiite a bit while in education, you couldn't live a full life yet they might have counted you as 'recovered'.

However, if we disregard the recovery label, the Nijhof study still shows substantial improvements in fatigue, school attendance and physical function. So even if no patients had fully recovered, and allowing for some self-report bias, a good number of patiets were almost certainly leading more active lives. With no cure on offer, that's not a bad thing.

My big concern about this study is that the results might be a flash in the pan. Previous CBT studies have shown dramatic 'recovery' figures eg one by Wessely, I think (I'm sure you'll know) and that Knoop paper too. Yet the control group in this study, (66% CBT?) showed lousy results. They really do need to replicate these results. And if they do, as well as offering internet CBT, they will proably need to close down their normal CBT/GET services as woefully ineffective.

 

[Cort]

Let's be honest, if these kind of results were found in a study that was looking at bio-medical causes we would stand up and take notice (as has happened with the Rituximab trial). This seems to be far more clear cut than the PACE trial, whatever the cause of this illness is, it seems that there is some subset of people who suffer with idiopathic chronic fatigue for whom this kind of intervention is effective. I feel both sides of the argument are culpable in preventing progress, because one side is so adamant that the cause is not psychological that they flatly dismiss any finding which would implicate that hypothesis, but the other side doesn't even acknowledge that there may be a spectrum of different illnesses, and that applying the results of studies done on patients with idiopathic fatigue to people with chronic ME/CFS is misleading and utterly useless. I think the future is very bleak indeed if we can't separate and define these illnesses properly.

We always assume that if CBT or another behavioral treatment helps that it means this disorder is psychological. But what if the Lights theory is correct - that the sensory nerves are blasting the brain with signals - sending pain and sensation levels upwards - with every activity? And/or what if the autonomic nervous system - which controls blood flows and regulates the immune system and is an important part of the stress response is dysfunctional?

Then you have a stress response that is on all the time and you could have loads of pain signals to the brain everytime it is activated. If that's true the stress reduction therapies can be helpful. I just wished they worked better

I think I know where I fit - I'm in the screwed up stress response group. Others may be in the pathogen fluey completely wiped out group.

 

[Jenny]

I think I know where I fit - I'm in the screwed up stress response group. Others may be in the pathogen fluey completely wiped out group.

I'm in both.

 

[Enid]

Whatever these people claim the science now revealing pathologies/infection/viral/bacterials involved in so many illnesses like ME place them as utterly secondary to biology.

Encouragement in whatever form is not going to solve ME.

 

[Guido den Broeder]

Let's be honest, if these kind of results were found in a study that was looking at bio-medical causes we would stand up and take notice (as has happened with the Rituximab trial). This seems to be far more clear cut than the PACE trial, whatever the cause of this illness is, it seems that there is some subset of people who suffer with idiopathic chronic fatigue for whom this kind of intervention is effective. (...)

Remember that there is no proper control group in this study. We don't know how many would have recovered without any treatment whatsoever.

I they do what they always do, Nijmegen will now ask for additional funding to include a no-treatment group, still overlook some other triviality (say objective measurement) in that new study, then ask for more funding again, etc., until finally they will use FITNET as 'usual care' and start a new cycle with another form of CBT treatment.

With this study, I say focus first of all on the extremely poor results of 'usual care'. Note that it is the same treatment that, for many years, Nijmegen claimed to produce a 70% recovery rate!

 

[Vitalic]

We always assume that if CBT or another behavioral treatment helps that it means this disorder is psychological. But what if the Lights theory is correct - that the sensory nerves are blasting the brain with signals - sending pain and sensation levels upwards - with every activity? And/or what if the autonomic nervous system - which controls blood flows and regulates the immune system and is an important part of the stress response is dysfunctional?

Because despite the rather shoddy definitions of recovery used in many of these studies it is undeniably the case that people do experience a full recovery with these talking therapies on a regular basis (be it CBT, MT, LP etc.), if there was an underlying neurological abnormality as the Lights postulate then it seems that the best these techniques could hope to achieve is a reduction in symptoms, as reducing stress via altering thought patterns would only be masking the underlying physiological issue. I suspect the primary cause within a significant portion of the CFS spectrum really is psychological, or neuro-psychiatric to be more precise, and if that were the case we are wasting our time trying to incorporate that subset within ME, we would be better admitting that for that subset the psychologists are correct in their evaluation and to let them get on with it, while we get on with solving ME. I don't see any way forward while the different subsets remain conflated as they currently are. I should clarify that I wasn't trying to accuse people of bias, I appreciate the efforts people make here to analyse each study in meticulous detail.

 

[Snow Leopard]

it is undeniably the case that people do experience a full recovery with these talking therapies on a regular basis (be it CBT, MT, LP etc.)

I disagree, the evidence isn't there, even in the studies that used loose definitions of CFS.

Evidence of recovery involves objective evidence of normal activity levels. In terms of employment outcomes over periods of 6-12 months, this has never been demonstrated.
Lots of studies of these interventions simply withheld this evidence because it did not show the 'expected' improvement desired by the authors.

Likewise, in terms of actigraphy this has never been demonstrated. CBT works in the sense that people complain less about their symptoms and greatly overestimate their level of function (in the short term - the effect disappears after a few years).

 

[Esther12]

Because despite the rather shoddy definitions of recovery used in many of these studies it is undeniably the case that people do experience a full recovery with these talking therapies on a regular basis (be it CBT, MT, LP etc.), if there was an underlying neurological abnormality as the Lights postulate then it seems that the best these techniques could hope to achieve is a reduction in symptoms, as reducing stress via altering thought patterns would only be masking the underlying physiological issue. I suspect the primary cause within a significant portion of the CFS spectrum really is psychological, or neuro-psychiatric to be more precise, and if that were the case we are wasting our time trying to incorporate that subset within ME, we would be better admitting that for that subset the psychologists are correct in their evaluation and to let them get on with it, while we get on with solving ME. I don't see any way forward while the different subsets remain conflated as they currently are. I should clarify that I wasn't trying to accuse people of bias, I appreciate the efforts people make here to analyse each study in meticulous detail.

This was pretty much the position I held prior to doing a lot of reading around the PACE trial. The results for that were so poor, and the way the results were spun so dishonest, that it has left me think that those promoting psychosocial treatments for CFS could just be 100% wrong, and totally deluding themselves about their worth and expertise.

But then... the results from this study look far more significant.

But then... we have had similar results before, only for it to later emerge that objective measures of activity and neuropsychaitric performance showed no benefit from CBT, but were not included in the initial papers purporting positive affects. The politicised nature of research around CFS makes it difficult to say what's going on.

In addition to all that, in theoretical terms, I'd have thought that psychiatric approaches could 'cure' physical causes of fatiguing illness (excuse the dualism), if they involved encouraging patients to adopt more simplistic and stupid views of reality and their lives and responsibilities. Thinking honestly about life takes up a fair amount of our energy, so even if someone has some problem with their immune system/muscles/brain injury then shifting resources away from an honest engagement with reality and towards what the therapist considers to be more 'functional' activities could lead to significant 'improvements'. So long as you take a rather disdainful view of patients. Such an approach to illness would be utterly unacceptable for me, but given the low views many have of those diagnosed with CFS, it wouldn't surprise me if many doctors saw this as a wonderfully pragmatic response to fatigue. The KCL website used to talk about their patients adopting a more 'spiritual' approach to life, and in my (limited and prejudiced) experience those who have benefited most from CBT for CFS seem to have adopted a rather glazed view of life... but the poor results from PACE would seem to indicate that they're not able to get much benefit from any trade-off here if that is what's occuring, and all of the above is just wild speculation really.

(edit: None of the above is to say that those with CFS cannot genuinely benefit from reasonably done CBT - 'CBT' can be performed in all manner of different ways, and it is only certain approaches which seem to promote self-delusion and dishonesty. The tolerance for such pragmatism in CFS does make it harder for patients to trust and access any sort of psychological help though).

Bit of an OT ramble there.