Published in Behavioural Sciences, 11 August 2017.
In November last year, when it was announced that the UK would be conducting a £1.1million internet-based trial of cognitive behavioural therapy among adolescents, it was a previously published Dutch trial that provided many of the media sound-bites.
The Dutch trial had claimed a 63% recovery rate that was maintained at follow-up – less than three years after the trial had completed. Usual care had similar rates of recovery at follow-up but these were achieved at a slower pace.
|‘The FITNET-NHS study is a randomised controlled trial comparing two treatments for children with Chronic Fatigue Syndrome/Myalgic Encephalomyelitis (CFS/ME) who do not have access to a local specialist CFS/ME service. The study will investigate whether FITNET-NHS, an online Cognitive Behavioural Therapy (CBT) program, is effective in the NHS, and whether it offers value for money compared to Activity Management (delivered via Skype).’|
There had already been concern expressed about the UK trial and the media coverage that accompanied it from Forward ME – the collaborative group that includes the ME Association and is chaired by the Countess of Mar.
And it is FITNET-NHS, and not biomedical research, that is the reason indicated by NICE (National Institute for Health and Care Excellence) recently, to remove the ME/CFS clinical guideline from the static list permitting an update to its content following publication of the trial.
Therefore it is important to consider just how effective the Dutch FITNET trial was and what this portends for the UK version. Dr Mark Vink and Dr Simin Ghatineh – both of whom are affected by M.E. – provide a review of both trials in their article, reproduced below.
FITNET’s Internet-Based Cognitive Behavioural Therapy Is Ineffective and May Impede Natural Recovery in Adolescents with Myalgic Encephalomyelitis/Chronic Fatigue Syndrome. A Review
The full review article is open access, but for the purposes of this ME Association blog we have taken some pertinent extracts by way of summary and these appear below:
The Dutch Fatigue In Teenagers on the interNET (FITNET) study claimed that after 6 months, internet based cognitive behaviour therapy in adolescents with Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS), led to a 63% recovery rate compared to 8% after usual care, and that this was maintained at long term follow up (LTFU).
Our reanalysis shows that their post-hoc definition of recovery included the severely ill, the unblinded trial had no adequate control group and it used lax selection criteria as well as outcomes assessed via questionnaires rather than objective outcomes, further contributing to exaggerated recovery figures.
Their decision not to publish the actometer results might suggest that these did not back their recovery claims. Despite these bias creating methodological faults, the trial still found no significant difference in recovery rates (“~60%”) at LTFU, the trial’s primary goal.
This is similar to or worse than the documented 54–94% spontaneous recovery rates within 3–4 years, suggesting that both FITNET and usual care (consisting of cognitive behaviour and graded exercise therapies) are ineffective and might even impede natural recovery in adolescents with ME/CFS.
This has implications for the upcoming costly NHS FITNET trial which is a blueprint of the Dutch study, exposing it to similar biases.
‘Myalgic Encephalomyelitis (ME) also known as Chronic Fatigue Syndrome (CFS) or ME/CFS
is a debilitating multisystem disease as concluded by the American Institute of Medicine (IOM) in
February 2015 after an extensive review of the literature.'
‘It is characterized by rapid (muscle) fatigability after trivial exertion with an abnormally delayed recovery, pain, neurological complaints and autonomic and immunological dysfunction, often resulting in severe functional limitations rendering at least 25% of patients house and/or bed bound.'
‘The IOM also concluded that there is no effective treatment for it and the American Federal Agency for Healthcare Research and Quality, removed its recommendation for CBT and GET in July 2016, after concluding that there is no evidence that these treatments are effective.'
‘Patient surveys from ME associations over the last few decades including three recent ones from the British, Norwegian and Dutch involving more than 3000 patients, and also a survey of young ME patients, have shown that there is no effective treatment/cure for this disease.'
‘All this is in contrast with the conclusion of the Dutch FITNET trial that 60% of participants recovered after treatment with internet-based CBT. This has led to the NHS allocating almost £1 million of funding, to replicate the Dutch study in the UK.'
‘In this article we re-analysed the FITNET trial, in an attempt to find out whether internet based CBT can lead to actual recovery from CFS in children/adolescents and whether there were any other important findings and if so, what they were.'
Summary of the Dutch FITNET Trial
‘The FITNET (Fatigue In Teenagers on the interNET) trial was registered as a randomised controlled trial involving 135 adolescents with CFS aged 12–18 years at the trial’s start.'
‘It compared the effectiveness of internet-based CBT with that of usual care, as although face-to-face CBT is reported to be effective in two-thirds of adolescents with CFS, it requires specialized therapeutic skills that are not always available within driving distance.'
‘The trial’s primary goal was the recovery rate at long-term follow up (LTFU). Primary outcomes were fatigue severity, physical functioning and school attendance assessed via questionnaires.'
‘The trial concluded that after six months, i.e., at end of treatment, 60% of participants had recovered in the Internet based CBT group and only 8% in the usual care group.'
‘At LTFU (mean of 2.7 years after treatment’s start) the positive effects of FITNET were maintained, 58.9% of adolescents had recovered from CFS. Usual care led to similar recovery rates, although these rates were achieved at a slower pace.'
Strengths and Weaknesses of the Dutch FITNET Trial
‘A particular strength of the trial was that they did not rely exclusively on subjective outcomes but also used the actometer, an objective measure of activity.'
‘The second strength of the study was their primary goal of recovery at long-term follow-up, which removes the effects of the naturally occurring fluctuations of CFS, according to the systemic review by Whiting et al.'
‘Its weaknesses were:
- a protocol that was published years after the trial started plus a badly designed control group;
- patient selection issues such as more than 70% of those in the treatment group not having an infectious onset;
- not excluding participants with a comorbid psychiatric disorder or those who during the trial were found not to have CFS but a psychiatric disorder instead;
- a post hoc definition of recovery that included the severely fatigued/ill, yet labelling it as a strength of the study;
- not publishing the actometer results, thus causing reporting bias;
- not taking into account the naturally occurring recovery rates in adolescents with CFS as documented by many studies;
- ignoring its own primary goal (recovery rates at LTFU) as well as the null effect at LTFU.'
Issues with the NHS FITNET Trial
6.1. Background Information
‘According to the protocol “There is good evidence that CBT is effective” and “In particular, the PACE trial showed that both CBT and GET were…effective” (pp. 6, 7), whereas independent reviews of the PACE trial showed that these treatments are not effective.'
‘The protocol further states that “Those children with 3 months of disabling fatigue plus one symptom (Nice guidance) will be eligible. The NICE guidance will be used and not the Centre Disease Control (CDC) criteria, based on the relevance of NICE criteria to the NHS” (p. 11).'
‘Yet according to the NICE guidelines, the fatigue needs to include all of the following features: “new or had a specific onset (that is, it is not lifelong); persistent and/or recurrent; unexplained by other conditions; has resulted in a substantial reduction in activity level; characterised by post-exertional malaise and/or fatigue (typically delayed, for example by at least 24 h, with slow recovery over several days)” (pp. 14–15).'
‘The NICE guidelines also require patients to have at least one of ten listed symptoms.'
6.2. Protocol Design, Diagnosis and the Control Group
‘Please note that some changes were made to version 3.0 of the NHS FITNET protocol after we had submitted our paper for publication. One of these changes, namely the inclusion of post exertional malaise (PEM) as a compulsory diagnostic requirement in the latest version of the protocol, alleviates some of our concerns about patient selection. However it is unclear why this protocol change was made at this stage or why PEM wasn’t included in the original protocol in accordance with the NICE guidelines being used.'
‘According to the NHS FITNET protocol, participants assigned to internet-based CBT will receive 19 sessions lasting 60 min over a six-month period while those in the activity management group “will have up to three video (e.g., Skype) appointments (one assessment and two follow up)” from an occupational therapist from the trial (after which the participant’s nominated local therapist or doctor is asked for a “review within six to eight weeks”) (p. 13).'
‘Yet in evaluating a psychological treatment, it’s important that the control group offers the same frequency and intensity of contact, positive expectations, attention and support as the intervention group otherwise any difference between the two groups might be simply down to the design of the study leading to the erroneous inference of efficacy in its absence.'
‘The protocol doesn’t state the maximum entry score for their primary outcome nor does it define improvement or recovery.'
‘Consequently, this trial, like other CBT trials is an unblinded trial by definition which relies on one subjective outcome. Yet as mentioned above, unblinded trials should use objective primary outcomes (instead or as well), to safeguard against erroneous inference of efficacy in its absence.'
6.3. Illness Fluctuations
‘The authors themselves state that: “CFS/ME is by its nature, a fluctuating illness” (p. 22), yet their primary outcome (disability measured using the Physical Function Scale (SF-36-PFS)) is directly at the end of treatment (at 6 months) and not as a systematic review by Whiting et al. advised, at least 6–12 months after a trial ended, to eliminate the effects of these fluctuations.'
6.4. Activity Management
‘The protocol states that “Activity Management is used as the comparator in this study as it is recommended by the National Institute of Health & Care Excellence (NICE)” (p. 5).'
‘[However,] …it seems that the NHS FITNET trial isn’t using activity management but a form of GET by instructing participants to increase activity by “10–20% each week” (p. 14), irrespective of symptoms.'
6.5. The Boom and Bust Theory
‘The NHS FITNET trial authors are using their ‘boom and bust’ theory of CFS to blame any deterioration on the disease, and not on the adverse effects of their own treatment.'
‘This was disproved by Van der Werf et al. (co-authored by an investigator also involved in both the Dutch and the UK NHS FITNET trial) and in a systemic review by Evering et al.'
‘Using or re-analysing actometer readings, both found no significant differences during the day or in the day-to-day fluctuations in activity patterns between controls and CFS patients.'
7. Dutch FITNET Discussion
‘There were a number of important issues with the trial including selection of the patients and the design of the control group.'
‘For instance, 72% of participants in the intervention group did not have a preceding viral illness/infection casting doubt on the validity of their CFS diagnosis.'
‘Also patients in the intervention group received 21 sessions of internet-based CBT, yet the exact nature of the treatment nor its intensity in the usual care group were specified to the point that 10% of participants in this group did not receive any treatment at all, ignoring the fundamental tenet of proper trial design that participants in the control group receive the same level of treatment, care and attention as those in the treatment group.'
‘The differences at the end of treatment might also be due to response shift bias or patients filling in questionnaires in such a way as to please the investigators. This is why rheumatology trials stopped the use of questionnaires in favour of using objective outcomes.'
‘The FITNET trial did not publish the actometer results. This is a form of outcome reporting bias that endangers the validity of a trial. Reanalysis of three other trials all involving one of the FITNET trial investigators, that did not publish their actometer results either, showed that CBT did not lead to objective improvement.'
‘In summary, regarding outcome differences at 6 months, it’s highly likely that the combination of an unblinded trial, subjective outcomes and a poorly chosen control group, alone or together with response shift bias and/or patients filling in questionnaires in a manner to please the investigators produced the appearance of positive effects, despite the lack of any substantial benefit to the patients, leading to the erroneous inference of efficacy in its absence.'
‘Furthermore, there was no difference in recovery rates between the two groups at LTFU, the trial’s primary goal. Crossing over from internet-based CBT to usual care or vice versa, did not influence recovery rates.'
‘The trial did not take into account the widely reported spontaneous recovery in adolescents with CFS (Table 1), which is similar to or better than the reported recovery rates in this trial.'
‘That is without taking into account the problems with the post hoc definition of recovery used by the FITNET trial which was tailor-made to fit the results yet was so wide that it included the severely fatigued/ill.'
‘The trial acknowledged that had they used a definition of recovery with one standard deviation instead of two, their recovery rate would drop from 63% to 36%.'
‘But even this definition didn’t resemble actual recovery as defined by Kennedy, where symptoms are gone and patient health returns to pre-illness levels.'
‘This suggests that Internet based CBT and also usual care consisting of individual or group based rehabilitation programmes, CBT or graded exercise treatment, or both, might actually impede the naturally occurring recovery process in adolescents with CFS.'
‘Moreover, up to 35% of FITNET patients (similar to the 36% if only one standard deviation had been used), had a comorbid depression and/or anxiety disorder for which CBT is the most effective treatment according to a meta-analysis by Tolin et al.; confounding the results of a trial that used a behavioural treatment.'
‘Reeves et al., co-authored by one of the FITNET investigators, concluded in 2003 that these patients should be excluded from CFS research studies.'
‘The UK NHS has allocated almost a million pounds to replicate the Dutch FITNET trial based on the reported recovery rates at the end of treatment.'
‘It is likely that had they been aware of the above issues of the Dutch FITNET trial, they might have reconsidered this.'
‘As the 2 trials are almost identical, the current NHS trial will also suffer from the above-mentioned methodological problems of the Dutch study.'
‘In conclusion, it is paramount that the general public as well as patients and health professionals, be aware of the proclamations of outcome success and high recovery rates from psychological interventions in the absence of objective data supporting these claims, especially when patients repeatedly testify that these treatments are ineffective and even harmful.'
‘Our reanalysis of the FITNET trial showed that this trial suffered from severe methodological problems including using a definition of recovery that included the severely ill.'
You can read the full review here.
Image copyright: bialasiewicz/123RF Stock Photo