By David Tuller, DrPH
Last year, BMJ Open published a paper called Randomised controlled trial of online continuing education for health professionals to improve the management of chronic fatigue syndrome: a study protocol. The seven authors, all affiliated with the University of New South Wales in Sydney, included Professor Andrew Lloyd, the infectious disease expert and the country€™s leading proponent for the GET/CBT rehabilitative approach to the illness he prefers to call CFS.
Professor Lloyd was the second-to-last author, not the first or senior (and corresponding) author. But the online continuing education program described in the protocol reflects the approach he has pioneered at his Sydney fatigue clinic. This online program is designed for allied health professionals, physiotherapists, exercise physiologists, psychologists, and others outside the domains of medicine. The program promotes graded exercise therapy and cognitive behavior therapy as evidence-based treatments for the illness.
The protocol calls for the trial to recruit 180 participants, divided into those who receive the online continuing education program and those who don€™t. (Those who don€™t will be offered the program afterwards.) Proposed outcomes of the study are how well the participants have absorbed this information and how comfortable they now feel recommending and delivering such treatments in practice. Hm.
Part of the context for this effort to disseminate information about these interventions is that Australian patients find it hard to obtain government and private disability support and payments without having been through a rehabilitative program. Since these two treatments are considered the standard of care in Australia, claimants are expected to have pursued either or both in their efforts to get well. Making such services more widely available would presumably allow more patients to obtain needed benefits–if not having undergone GET or CBT remains an obstacle.
Another solution, of course, would be to stop making such payments dependent on having undergone these therapies. But so far, government agencies as well as the Royal Australian College of General Practitioners seem not to have noticed that international support for GET and CBT as the standards of care for ME/CFS has been disintegrating. In the U.S., the Centers for Disease Control removed the therapies from its recommendations for the illness last summer. Australian patients and advocates have tried to alert government and medical authorities to these developments but so far have been largely ignored.
BMJ Open received the protocol manuscript on September 2, 2016. A revised version was submitted on March 7, 2017, and accepted a week later. The public challenges to the GET/CBT paradigm and to PACE specifically were well underway by this point. Virology Blog posted an open letter to The Lancet outlining the PACE trial€™s unacceptable flaws a year before–in February, 2016. The anonymous trial data were released in September, 2016, and the Journal of Health Psychology published Keith Geraghty€™s PACE-Gate editorial that November.
Yet the BMJ Open protocol for the online continuing education program does not discuss the issues that have been raised for years about the validity or reliability of the evidence for GET and CBT, beyond an opaque reference to the controversy surrounding the PACE trial. The stated justification for the online program is that allied health professionals have too little knowledge or understanding of these interventions and/or don€™t feel sufficiently equipped to recommend or implement them appropriately.
The protocol and the online continuing education program therefore rest on the assumption that the evidence base for CBT and GET is solid. I don€™t share that assumption. I therefore don€™t have a high opinion of the protocol or the online program it describes involving these purportedly evidence-based interventions.
The protocol€™s opening lines generate immediate concerns. Here€™s the first sentence: Chronic fatigue syndrome (CFS) refers to the presence of persistent and severe fatigue that is accompanied by musculoskeletal pain, neurocognitive difficulties, in addition to sleep and mood disturbances, and cannot be accounted for by a medical condition. To support this statement, the authors cite the CDC€™s 1994 Fukuda case definition.
This is problematic. The written text does not match the Fukuda definition. Besides fatigue, Fukuda requires four of eight additional symptoms. That means Fukuda allows for, but does not specifically require, musculoskeletal pain, neurocognitive difficulties or sleep disturbances. Moreover, mood disturbances, which presumably refers to depression and anxiety disorders–are not among Fukuda€™s set of eight symptoms.
This mismatch between text and citation suggests the authors don€™t much care how chronic fatigue syndrome is defined, as long as whatever is being investigated has been called chronic fatigue syndrome. Two subsequent sentences–which cite the Cochrane reviews and more than 20 randomised controlled trials in support of GET and CBT–reinforce that impression. How? Because these reviews and publications include many studies relying on the Oxford case definition, which only requires one symptom–six months of unexplained, debilitating fatigue.
This case definition generates samples that are more likely to include people with predominantly psychiatric disorders than more stringent case definitions, and people with depression and anxiety are known to respond positively to exercise and psychotherapy. Moreover, patients who meet the Oxford criteria because they suffer from fatigue for whatever reason might not experience what has been recognized as the cardinal symptom of ME/CFS, post-exertional malaise, or what the U.S. Institute of Medicine (now the National Academy of Medicine) referred to as exertion intolerance in a major 2015 report.
In other words, patients who suffer from the actual disorder can experience severe relapses after minimal levels of activity. In such circumstances, a program focused on steady increases in activity, like GET, could be considered contraindicated. In multiple surveys, more patients have indicated that they were harmed by a GET program than that they were helped. (To be fair, the protocol authors warn of possible post-exertional exacerbation if GET is improperly delivered. Given that, it is not clear why they accept evidence from studies using the Oxford criteria, which does not require the presence of post-exertional malaise as a symptom.)
It is a questionable enterprise to extrapolate findings from a population selected through broad criteria to those defined more narrowly. It bears repeating that the heterogeneous nature of samples derived using the Oxford criteria led a 2015 report from the National Institutes of Health to determine that this case definition could harm research and should be retired. That suggests that Oxford studies should also be retired from use as the basis of treatment recommendations.
When the U.S. Agency for Healthcare Research and Quality reanalyzed data on these interventions but removed Oxford studies from the sample, it found little to no reason to recommend CBT or GET. Cochrane so far has not conducted a similar reanalysis. For that and other reasons, its reviews of GET and CBT for ME/CFS cannot be taken at face value. The same can be said of many of the more than 20 studies cited in the protocol as further evidence of the effectiveness of GET and CBT.
As the CDC presumably recognized in dropping the GET/CBT recommendations, these case definition issues have implications for interpreting research findings, but the protocol for the online continuing education program does not appear to address them. Beyond this concern, the bulk of the GET and CBT studies suffer from a major weakness: they are open-label trials relying on subjective outcomes. This study design is at high risk for generating biased responses.
In contrast, objective measures have largely failed to confirm the self-reported benefits of GET and CBT. All the objective outcomes in PACE produced poor results. Those receiving the treatments were not more fit, more likely to get back to work or more likely to get off benefits. Moreover, at least three other studies used monitors to track how much participants moved and found no benefits in the treatment arms, undermining the subjective reports of improvement.
Will the health professionals in the online continuing education program be informed about the null objective results for the rehabilitative treatments being promoted? Will they learn about the problems with relying on subjective outcomes in open-label trials? Will they be told that the CDC no longer considers GET and CBT to be evidence-based treatments for this illness? Perhaps, but the protocol is silent on these matters.
To provide health professionals with an online continuing education program that overlooks the many concerns leading scientists have raised not only about PACE but the entire GET/CBT paradigm is to do them, and their prospective patients, a disservice. The authors of this protocol appear to have convinced themselves that the evidence base for GET and CBT is beyond serious challenge. They have also convinced themselves that any concerns can be allayed with a casual mention of the controversy over the PACE trial. They are wrong.
On a related issue, the protocol of this online continuing education program notes the absence€¦of curative treatments and refers to the interventions as designed for managing symptoms. This suggests that the therapies are being promoted on a different theoretical basis than the body of evidence cited to support their effectiveness. In PACE, the illness is presented as perpetuated solely by deconditioning and therefore reversible with increases in activity. The descriptions of the interventions are clear: They are based on the hypothesis that there is no organic cause of ongoing illness and that recovery is therefore possible.
The protocol for the online program makes no mention of this deconditioning hypothesis and the notion that the illness is reversible. Nor does the protocol offer a new rationale for these therapies. However, one of the articles cited by the protocol, a 2016 study of outcomes from Professor Lloyd€™s fatigue clinic, provides an insight.
In the description of GET in this clinic-based study, Professor Lloyd and his co-authors explain their hypothesis: The GET component was based on the notion of central sensitisation to physiological signals and hence desensitisation (i.e. carefully graduated exposure to the triggering stimuli–exercise) to reduce the exaggerated (brain-derived) symptom of fatigue.”
This assertion certainly represents a shift from the deconditioning framework that has dominated the research into rehabilitative treatments. The authors of a 1989 article, including both Simon Wessely and Trudie Chalder, described a pattern of sedentary behavior and mistaken illness beliefs leading to a chronic fatigue state in which symptoms are perpetuated by cycle of inactivity, deterioration in exercise tolerance and further symptoms. The GET and CBT recommended for what was being called the chronic fatigue syndrome arose out of this hypothesis.
Within this framework, the GET was intended to re-condition patients. The CBT was geared toward the same end by seeking to alleviate patients of the unhelpful conviction that an ongoing organic disease prevented them from engaging in activity. Unlike the CBT often offered for patients with cancer, MS or other recognized diseases, this intervention was not designed to help people adjust to their terrible condition but to hep them overcome the false cognitions believed to be contributing to the perpetuation of their illness. The GET/CBT ideological brigades have championed this deconditioning hypothesis for years. It reached its apotheosis in the PACE trial.
The switch to the central sensitization theory is a relatively recent development. It is of course possible for therapies to work even when the underlying mechanism is not understood, or when ideas about that underlying mechanism shift. But if all the evidence is based on one set of claims and conditions, it would be helpful if those substituting a different set of claims and conditions provided a reasonable and thorough scientific explanation for that decision. That is especially the case when the other theory is as speculative as central sensitisation of brain-mediated fatigue pathways appears to be.
The protocol for the online continuing education program does not address this issue of hypothesis-switching, although the issue seems germane to the continuing education being proposed. If the deconditioning hypothesis that has anchored the GET/CBT paradigm since 1989 is no longer operative, those who seek to replace it with something else should explain why they have abandoned it, and why their new hypothesis is more likely to remain viable. They should also explain what this change means for the credibility and relevance of PACE and the body of GET/CBT research, in which these rehabilitative approaches were specifically designed to address the presumed central role of deconditioning in perpetuating the illness.
Here are some questions that the authors of this protocol for the online continuing education program should answer:
*What has happened to the deconditioning hypothesis, and why?
*Does this switch in hypothesis mean that CFS is no longer reversible and that recovery is no longer possible through GET and CBT?
*Does this switch in hypothesis require other changes in the way GET and CBT are conceptualized, structured, implemented, and delivered?
*How do such changes impact the validity and reliability of claims of effectiveness derived from studies in which GET and CBT were developed and tested based on the now-defunct deconditioning hypothesis?
*Does the overall failure of objective measures to match subjective reports of treatment success raise doubts about the reliability of the subjective reports?
*Given the minimal evidence backing up the central sensitization hypothesis itself, what is the evidence to support the idea of de-sensitization through exercise?