Over the past 5 years, several reviews of clinical interventions for ME/CFS were published by prestigious medical institutions. These include the report by the National Institute for Health and Care Excellence (NICE) in the UK, the review commissioned by the Centers for Disease Control and Prevention (CDC) in the US, the reports of the Health Councils of Belgium and the Netherlands, and as of this week, the report by the German Institute for Quality and Efficiency in Health Care (IQWiG).
These reports present comparable findings and appear to establish a new consensus on the treatment of ME/CFS. They deviate from previous guidelines that endorsed Graded Exercise Therapy (GET) and Cognitive Behavioral Therapy (CBT) as evidence-based treatments. The new reviews critically evaluate these therapies, incorporating methodological critiques and reports of harm. They focus on recent diagnostic criteria for ME/CFS where post-exertional malaise (PEM) and not fatigue is considered the most characteristic feature of the syndrome.
The new consensus no longer recommends GET but still includes conflicting and ambiguous statements about this controversial treatment. It is more favorable about CBT but assumes it is a supportive treatment that aims to help patients cope with their illness. Both conclusions are a step in in the right direction but appear to be a diplomatic compromise rather than an accurate scientific evaluation of the evidence. I will try to illustrate this by focusing on a few oddities that these new guidelines have in common.
Bias ceiling and checklist methodology
The first is what I would call the ‘bias ceiling’ effect for lack of a better term. All the reports have difficulty stating that randomized trials on GET and CBT have critical flaws that make their results unreliable.
They used different tools to assess study limitations and risk of bias. The NICE guideline, for example, used GRADE while the CDC review used the methods guide of the Agency for Healthcare Research and Quality (AHRQ). The IQWiG has its own rules for assessing evidence but these methods are all quite similar. They seem to stop considering the impact of study limitations after a certain point.
Take for example the recent German report by IQWiG. It noticed that trials on GET and CBT are at high risk of bias because they relied on subjective outcomes while participants and therapists were not blinded. Because this was the highest risk of bias in their classification system, it was rather useless to look for additional flaws. Page 124 of the report states (translated from German): “Since the risk of bias of all 3 studies included was already classified as high across outcomes, there was no outcome-specific assessment of the risk of bias.”
GRADE classifies certainty of evidence into 4 categories: ‘High’, ‘Moderate’, ‘Low’, or ‘Very Low’. Study limitations allow downgrading the certainty of evidence but only with 2 levels. This is a problem because randomized trials start out as having high-quality evidence. Hence if you detect numerous serious problems (and thus a very high risk of bias), the certainty of evidence would still not be classified as ‘very low’. There is no good justification for restricting the impact of study limitations to only 2 levels of certainty. What if there is a flaw that makes the study results useless? This seems to be a dead angle in the GRADE system.
The CDC review had similar problems. The authors downgraded the evidence on GET and CBT with only one level because, in their view, that is the standard recommendation for open-label trials where blinding of patients and therapists was not possible. They appear to downgrade the quality of evidence with a fixed amount because this is specified in a manual rather than think critically about how the flaw might distort the results in particular trials. Both GET and CBT aimed to influence how patients interpret and report their symptoms, meaning that there was arguably a high risk that questionnaire results were biased.
This ‘checklist methodology’ of assessing bias has made reviewers malleable to researchers who repeatedly use the same flawed trial design to promote their theory. It also explains why it has been difficult for reviews on ME/CFS to state the obvious, namely that the evidence on GET and CBT is not reliable. Most keep on suggesting that there is limited but credible evidence supporting these therapies despite the many study limitations and high risk of bias. Their methodology overlooks the possibility of a critical flaw, one that does not downgrade the quality of evidence with 1 or 2 levels but makes it entirely useless. Even a small hole can make a ship sink.
Most of the reports preferred not to recommend GET as a treatment for ME/CFS. The evidence for it is clearly limited and many patients, physicians, and researchers in the field are vehemently opposed to the intervention. It therefore seems that the reviewers used additional reasons for not recommending GET such as the outdated diagnostic criteria, statistical heterogeneity, reports of harm, or simply the controversy about this treatment. Some of these reasons are valid but might not be persuasive to medical professionals outside the ME/CFS community. They give the impression that no recommendation was made because this is an area of controversy and conflicting views.
More correct would be to state that the evidence supporting GET is simply not reliable enough. This should not need additional arguments. The guidelines could, for example, have restricted their recommendations to trials that have a low risk of bias. The CDC review only found 4, none of which were about GET or CBT. Things would have been made clear in only a few pages, without controversy.
Inventing CBT: have your cake and eat it
The conclusion on CBT is even more bizarre.
All reviews suggest that there is scientific evidence that CBT is helpful for ME/CFS patients. But when they define CBT, they present it as a supportive treatment that helps patients cope with their symptoms. The German report for example states (translated from German):
“This therapy should help to better deal with the mental stress caused by the disease. In this way, strategies are to be learned that help in dealing with the symptoms and with consequences such as depressive thoughts and fears.”
One can find similar statements in other reports, but they do not correspond to how CBT was used in the randomized trials. In the trials, researchers used a psychosomatic model that assumes ME/CFS is reversible by addressing deconditioning, catastrophizing, and fear of activity. The largest trial, the infamous PACE study, said about CBT: “The aim of treatment was to change the behavioural and cognitive factors assumed to be responsible for perpetuation of the participant’s symptoms and disability”. In short: CBT aimed to treat or cure ME/CFS by increasing activity levels, not help with mental stress caused by the disease.
The NICE guideline in the UK also recommends CBT for ME/CFS patients. It includes the following paragraph:
“The committee wanted to highlight that cognitive behavioural therapy (CBT) has sometimes been assumed to be a cure for ME/CFS. However, it should only be offered to support people who live with ME/CFS to manage their symptoms, improve their functioning and reduce the distress associated with having a chronic illness.”
This sounds sensible but it contrasts with how CBT was used in scientific studies. In the trials, CBT was presented as a cure for ME/CFS. The research team of Gijs Bleijenberg, the Dutch team which published the most trials on CBT for ME/CFS, has been quite clear about this. They wrote:
“one of the last cognitive interventions in cognitive behaviour therapy is for patients to stop labelling themselves as CFS patients. In our opinion it helps them to strive for a cure, since a personal goal can never be completed if it is not aimed for. By questioning or denying the curative intentions of cognitive behaviour therapy, Clark and colleagues may deprive CFS patients from a potential cure.”
Similarly, British researchers have claimed that CBT helps patients recover from ME/CFS. They concluded that “recovery from CFS is possible, and that CBT and GET are the therapies most likely to lead to recovery.” There are almost no trials that used a different, supportive form of CBT. Pretty much all evidence on CBT summarized in the reviews is based on this psychosomatic version.
Reviewers were likely reluctant to endorse such a radical interpretation of CBT as this was causing problems and distress in the ME/CFS community. As a solution, they conceived their own interpretation of what CBT for ME/CFS should look like: a supportive therapy that helps patients cope with their symptoms. This is in line with how CBT is used for other physical illnesses. The problem, however, is that the reviews pretend that the evidence from randomized trials applies to the CBT version they made up instead of the CBT that Bleijenberg and the PACE researchers used.
You can’t have your cake and eat it.
The solution presented is a Belgian compromise, one that everyone can live with but that makes no sense at all. There are only two reasonable positions. Either you assume there is scientific evidence for the radical, psychosomatic version of CBT or you accept that there is currently no evidence on CBT for ME/CFS at all. And if there is no evidence, why should CBT be recommended?
Why is pacing recommended?
This brings us to the third point: pacing. This refers to a self-management strategy that helps ME/CFS patients stay within their energy limits and avoid post-exertional malaise. Many patients find pacing a useful strategy to avoid relapses, manage their symptoms and remain as active as their illness allows them to be. It is often something that comes naturally because the illness forces patients to pace their activities.
The IQWiG report, CDC website, and NICE guideline, however, present pacing as if it were a treatment option (they might use a different name for it such as ‘energy management’ or ‘activity management’, but the main principles are the same). In my view, these guidelines do not sufficiently make clear that there is no evidence that pacing alleviates ME/CFS symptoms. It has been tested in two studies, the PACE trial and a 2017 Norwegian trial by Pinxterhuis et al., and both found no effect compared to usual care.
The fact that pacing is still recommended in some of these new guidelines indicates that they are a compromise based on expert opinion rather than a rigorous evaluation of the scientific evidence.
Conclusion
In a relatively short period of time, several influential medical institutions have published reviews on the treatment of ME/CFS. They seem to form a new consensus, one where PEM is a required criterion for ME/CFS, where GET is no longer recommended and CBT is proposed as a supportive treatment. While these reports are a significant step forward for the ME/CFS community, they appear to be a diplomatic consensus or expert opinion rather than a rigorous assessment of the current evidence. It remains to be seen how this will influence the future of ME/CFS care. I thought some of these developments are rather curious and worth highlighting. I hope that his blog post will help patients understand the new reports and how they have been formed.
EDIT 18/05/2023: Someone noted that I could have made clearer that there is no evidence for any type of CBT, even the curative version used by the PACE researchers. He noted: “We therefore have the absurd situation where CBT proponents are trying to promote the effectiveness of their supportive type CBT by pretending that it’s the same as PACE-type CBT which has been tested endlessly without ever producing any reliable evidence that it is effective! […] It’s not so much that they want to have their cake and eat it. They are trying to have an imaginary cake which never existed (or went mouldy a long time ago) and eat it.” I very much agree with this statement. The situation is even more absurd that I sketched.
Like you, I have found it very frustrating that the evaluators of research don’t seem to understand how CBT is actually used. I wish they would use a different term, possibly Illness Management, for helping people to cope with the limitations and emotions relating to our illness and drop CBT altogether.
Great piece.
I don’t find recent developments surprising; medical institutions are always political institutions to some degree.
Hence, they will, as expected, react to political pressure. Whether that is good or bad is a different matter. At this point, it is a positive development for ME patients that might haunt us down the line but might help to save lives NOW.