• Phoenix Rising needs funds to operate: please consider donating to support PR

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

Marco

Grrrrrrr!
Messages
2,386
Likes
3,216
Location
Near Cognac, France
That would be interesting, but what I was really hoping for was something that looked at possible bias where trial participants had a strong relationship with a therapist, and might be, subconsciously or otherwise, boosting their score a little. I've not seen anything like that to date.
Here's the link and relevant extract.

http://www.cfids-cab.org/cfs-inform/Ptsd/mccue.etal06.pdf

Screening for psychological distress using internet administration of the Hospital Anxiety and Depression Scale (HADS) in individuals with chronic fatigue syndrome

However, the control group were observed to demonstrate significantly higher levels of HADS-assessed anxiety and depression compared with
normative control data (Crawford et al., 2001). This provides some support for Buchanans (2003) observation that Internet-based data collection (in the HADS and other instruments) may produce inflated scores due to the testing medium. It is possible that anonymous Internet administration results in increased self-disclosure or reduced socially desirable responding, and that Internet questionnaires may thus provide inflated estimates of levels of negative affect experienced by respondents. An alternative possibility is that the control group may represent a biased sample. Simply choosing to take part in an on-line personality test, which while not expressly reporting a formal diagnoses, may be suggestive of greater levels of underlying psychopathology.
This would suggest that internet delivery might reduce the effects of social desirability bias and encourage more frank disclosure. On the other hand its difficult to tell what other biases might be introduced. People behave strangely on the 'net'. The may also behave differently when someone has their e-mail address.
 
Messages
1,383
Likes
362
Location
UK
Here's the link and relevant extract.

http://www.cfids-cab.org/cfs-inform/Ptsd/mccue.etal06.pdf

Screening for psychological distress using internet administration of the Hospital Anxiety and Depression Scale (HADS) in individuals with chronic fatigue syndrome

This would suggest that internet delivery might reduce the effects of social desirability bias and encourage more frank disclosure. On the other hand its difficult to tell what other biases might be introduced. People behave strangely on the 'net'. The may also behave differently when someone has their e-mail address.
Thanks, Marco.

It may be that the assessment in this study was face to face:
Of the patients lost to follow-up (telephone interview),...and two who were not recovered wanted to continue the treatment but felt no need to come back for the postintervention assessment (both assigned to usual care).
@Dolphin. Thanks, I've now read the paper myself, and couldn't see any evidence in support of their approach either.
 
Messages
1,383
Likes
362
Location
UK
More on school attendance

I wonder if they simply wanted to turn up (or say they turned up) for those two weeks so they can be seen as compliant? Remember that children are much more erm, susceptible to suggestion than adults.
I don't know if this is relevant:
A particular strength is that
the main outcome (school attendance) was checked and
double checked by the investigators, parents, teachers,
and therapists.
There appears to be a surprisingly small difference between a 'functionally-impaired' school attendance of 85% or less and a 'recovered' school attendance of 90% or more:
functional impairment [was defined] as a score of 85 or less on the child health questionnaire (CHQ-CF87) physical functioning subscale, or a school attendance of 85% or less.

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

Dolphin

Senior Member
Messages
17,555
Likes
28,242
Id love to know what these children were actually being told in this CBT were they taught to be kind to theirselves and pace well so not in that constant push crash cycle so much which ends up affecting many of us and making us worst early ME/CFS??
Nothing close to that, I think it's fair to say.

The FITNET protocol is based on this:

Bleijenberg G, Prins J, Bazelmans E. Cognitive behavioral therapies.
In: Jason LA, Fennel PA, Taylor RR, eds. Handbook of chronic
fatigue syndrome. Hoboken: Wiley and Sons, 2003: 493526.
which can be read here: https://listserv.nodak.edu/cgi-bin/wa.exe?A2=ind1009C&L=CO-CURE&P=R1774&I=-3
 

Dolphin

Senior Member
Messages
17,555
Likes
28,242
School attendance is not too bad an objective measure. There is of course a problem that kids with mild ME could be encouraged to use all their energy going to school rather than doing other activities hence an increase in school attendance does not necessarily corrospond to an increase in activity.
I think that could be happening.
There was a lot of pressure on these to attend school:

(i) CBT therapist wanted it

(ii) Parents were doing a type of CBT and being convinced this is what the children should do

(iii) The children were prescribed a school monitor:
The FITNET therapist and school mentor had at least
one communication about school attendance and the
schools effort to encourage treatment compliance. The
school mentor acted as a coach, adviser, or tutor when
needed. School mentors were sent a standard letter at the
commencement of treatment asking them for their
cooperation and consideration.
(iv) the program in itself (what they had to read) would try to persuade the adolescent they should go to school.
 

Dolphin

Senior Member
Messages
17,555
Likes
28,242
Esther12 said:
I only had a quack (edit - genuine typo, 'quick') glance - but it looked like they talked of using actometers. That would be great if it got us some real data on how these different groups were responding to treatment.
I missed seeing that, you're right.

And guess what? They've cherry-picked the results so as to not include the actometer results in the Lancet publication. (and we know they've done that before)
Technically they can get away with it. It is listed more in the predictors section, "Possible patient's predictors of treatment outcome":
http://www.biomedcentral.com/1471-2377/11/23/table/T2
But they did use it at 6 months as 12 months as well as baseline.

alex3619 said:
Fatigue In Teenagers on the interNET - The FITNET Trial. A randomized clinical trial of web-based cognitive behavioural therapy for adolescents with chronic fatigue syndrome: study protocol: "(d) Physical performance as measured with the actometer. This is a motion-sensing device worn at the ankle that registers and quantifies physical activity. The actometer is worn day and night during a period of twelve consecutive days [27]."

Effectiveness of internet-based cognitive behavioural treatment for adolescents with chronic fatigue syndrome (FITNET): a randomised controlled trial: "Physical functioning was measured with the subscale physical functioning of the CHQ-CF87 (0100%). This assessment method is reported to be reliable and has been validated with a good internal consistency (Cronbachs ?=086)."

The PACE trial started with actometers in the design, but finished without them. Every study with actometers that I have read showed either no improvement in functional capacity or a decline in functional capacity.

This study dropped actometers from their protocol too. I wonder why?

Bye, Alex
As I mentioned above, we have no evidence that they never took the readings.
 

Dolphin

Senior Member
Messages
17,555
Likes
28,242
Measurement of School attendance


I agree we need to be aware of activity displacement to ensure overall activity levels have increased. However, I would caution against assuming actometers are a gold standard of measuring activity in 'free-living' humans - they are not. They are very accurate at measuring activity on a lab treadmill, so long as the treadmill isn't inclined up or down, in which case they are clueless as they measure acceleration not force/load. For example, an actometer can't tell the difference between strolling across the living room and climbing a flight of steps - wheras the difference is critical to a patient. The manufacturers of these devices have hyped the accuracy and this has largely been uncritically accepted by researchers. You can see an intermittent debate about this starting here.

Actometers are a valuable tool, and don't suffer from self-report bias, but we shouldn't assume they give a definitive value for activity levels. In particular, I have seen no evidence on their ability to accurately measure differences in activity levels pre/post intervention (and I haven't seen such evidence for SF-36 etc either).

Do you know of any specific references re survey bias? I know there's a lot of literature but never managed to find a particular killer paper; all the stuff I saw was tangential.
@user9876 I still think one can make points asking for actometers to be used. There are lots of different ones on the market. It's generally accepted they are better than questionnaires and I think Nijs et al recently referred to as them as the gold standard way to measure activity.
 

Dolphin

Senior Member
Messages
17,555
Likes
28,242
Protocol was submitted when the trial was almost finished

The Lancet paper says:
148 adolescents who were diagnosed with chronic
fatigue syndrome were screened for eligibility between
January, 2008, and February, 2010.
The protocol was received by the journal: 28 January 2011 (ref: http://www.biomedcentral.com/1471-2377/11/23).

Seems a bit odd.

It was registered so maybe they didn't change anything:
http://www.controlled-trials.com/ISRCTN59878666 and http://www.clinicaltrials.gov/ct2/show/NCT00893438
 
Messages
1,383
Likes
362
Location
UK
@user9876 I still think one can make points asking for actometers to be used. There are lots of different ones on the market. It's generally accepted they are better than questionnaires and I think Nijs et al recently referred to as them as the gold standard way to measure activity.
My point was that they are indeed referred to as a gold standard, but this is assumed rather than proven for 'free-living' humans. They give valuable information on activity and where actometers and questionnaires give confliciting answers that raises serious doubts about the validity of the findings - as happened in a previous CBT trial. So it's important to have the actometer data to help better gauge true activity levels and I agree it's worthwhile to ask for them to be used. I can't help feeling that if the actometer data strongly supported their findings, the authors would have mentioned it in this paper.
 
Messages
1,383
Likes
362
Location
UK

Dolphin

Senior Member
Messages
17,555
Likes
28,242
Error in Appendix (I'm pretty sure) - may or may not be important

Table 3b. Recovery at 6 months by treatment group (mean +/- 1 SD)

Fatigue severity (CIS)*

*CIS-fatigue improvement cut off score < 35
Mean +2 SD = 40.
This would mean that the SD is 5 and hence the mean is 30.
That's not right - what I have seen before is some number in the teens (possible scores are 8-56)
 
Messages
140
Likes
66
Location
SF Bay Area
School attendance is not too bad an objective measure. There is of course a problem that kids with mild ME could be encouraged to use all their energy going to school rather than doing other activities hence an increase in school attendance does not necessarily corrospond to an increase in activity.
Take a look at Table 4. In the last two rows, they have practically the same physical functioning score but a 25% difference in school attendance. Rows 3 and 7 have practically the same fatigue severity score but almost a 30% difference in school attendance.
 

Dolphin

Senior Member
Messages
17,555
Likes
28,242
Duration of illness at baseline

The authors refer to 2 years at one stage:
Our patients had symptoms for almost 2 years before chronic fatigue syndrome was diagnosed and treated,
However using median and range the figures are:

Duration of symptoms at entry (months)
FITNET (n=68) 160 (684)
Usual care (n=67) 190 (6108)
I'd expect adolescents who were ill for only that length of time to have a relatively good prognosis.
 

Sean

Senior Member
Messages
7,378
Likes
18,045
I'd expect adolescents who were ill for only that length of time to have a relatively good prognosis.
My recall is that one of the few predictors of natural improvement is duration of illness at diagnosis. An earlier diagnosis in an adolescent may be more likely because 1) they are much younger and have had less time to have the disease, and 2) may just be more closely monitored in general than adults (so any health trouble will get picked up earlier). If so then they will naturally tend to have better outcomes, independent of other factors. Throw in that adolescents (and children in general) seem to have better overall recoveries than adults for a wide range of medical problems...

That could account for a large chunk of the 'therapeutic' effect in this study.
 
Messages
1,383
Likes
362
Location
UK
Mean +2 SD = 40.
This would mean that the SD is 5 and hence the mean is 30.
That's not right - what I have seen before is some number in the teens (possible scores are 8-56)
Here's some workplace data on CIS norms: Fatigue among working people: validity of a questionnaire measure, see Table 2

White collar worker 20.3 (10.1) mean (SD)

Blue collar: 21.9 (11.4)

After hernia: 32.9 (11.7)

Pregnant: 33.9 (13.0)

Mental: 39.6 (12.1)

On these figures (which are from some of the FITNET co-authors), 30 does look high for a healthy mean. Though maybe adolescents are more fatigued than adults...

A nature paper on bone marrow transplation gives mean fatigue levels under 30 even though around 1/3 of theseare stated as having severe fatigue.

I've also seen >35 as the threshold for severe fatigue eg a Bleijenberg paper on Rheumatoid Arthritis.


Ah, maybe this is it: Is a Full Recovery Possible after Cognitive Behavioural Therapy for Chronic Fatigue Syndrome?
The CIS-fatigue indicates the level of experienced fatigue
over the past 2-week period and consists of 8 items on a 7-point
scale. The score can range between 8 and 56 [9] . A normal group
of 53 healthy adults with a mean age of 37.1 (SD 11.5) has a mean
score on the CIS-fatigue of 17.3
(SD 10.1). Using this as a reference
for the CBT group, there resulted a threshold score of 27, the mean
plus 1 SD [13] .

13 Vercoulen JHMM, Alberts M, Bleijenberg G:
De Checklist Individual Strength (CIS) (The
Checklist Individual Strength). Gedragstherapie (Behavioural Therapy) 1999; 32:
642649.
 

PhoenixDown

Senior Member
Messages
408
Likes
801
Location
UK
I'd like agree with earlier posters who pointed out that school attendance shows us very little of their overall performance, what about improvement in grades? How long were they at school for? Did they take all the classes that healthy kids took? This could be a case of forcing attendance without significant improvement of the academic disability that this illnesses can cause.

My recall is that one of the few predictors of natural improvement is duration of illness at diagnosis. An earlier diagnosis in an adolescent may be more likely because 1) they are much younger and have had less time to have the disease.
Disease, what disease is that then? What was the diagnostic criteria? Seems like they are letting anyone with 25% fatigue in to this illness category.

It also looks like they are talking about general fatigue alone, where was the mention of balance problems, severe memory problems, vocabulary regularly disappearing into thin air, auditory processing disorders, POTS, slurred speech, tremors, severe concentration issues?

It really is long over due for society to stop using this waste-basket term. This explains why I was once told by my doctor that "You can't be that ill, you only have CFS".